Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
27229573
© University of Pretoria
ABSTRACT
ABSTRACT
Airborne dust and methane are common problems in the underground coalmines.
They pose health and safety risk to mining personnel, and a safety risk to mining
equipment as well. In order to prevent these risks air borne dust and methane
concentrations must be reduced to within the acceptable levels. In South Africa, the
dust and methane concentration in coalmines should not exceed 2.0 mg/m³ and 0.5%
per volume, respectively.
Mine ventilation is one of the popular ways of controlling both dust and methane.
Different ventilation systems have been designed since the history of underground
coal mining. Unfortunately, none provides ultimate solution to the dust and methane
problem, especially in the most critical areas of the underground coalmine, like blind-
end of the heading and last through road.
i
ABSTRACT
After verification further numerical analysis was done to in order to device a method
for determining optimum fan positions for different heading dimensions.
This study proves that CFD can be used to model ventilation system of a scaled down
coalmine model. Therefore chances that this might be true for the actual mine are
very high but it needs to be investigated. If this is found to be true then CFD
modelling will be a very useful tool in coalmine ventilation system research and
development.
ii
ABBREVIATION
ABBREVIATIONS
AC Alternating Current
ARD Airborne Respirable Dust
CAD Computer Aided Drawing
CFD Computational Fluid Dynamics
CM Coal Miner
CPU Computer Processing Unit
CSIR Council for Scientific and Industrial Research
CWP Coal Workers’ Pneumoconiosis
DAU Data Acquisition Unit
DC Direct Current
DME Department of Minerals and Energy
FVM Finite Volume Method
GM Mine Geometry
KSSC Kloppersbos Shear Spray Curtain
LED Light Emitting Diode
LTR Last Through Road
PDE Partial Differential Equation
RSA Republic of South Africa
USA United States of America
iii
SYMBOLS
SYMBOLS
1
1 tonne = 1000 kg
iv
SYMBOLS
hL Head loss m
K Head loss coefficient
k Conduction of continuum W/mK
L fcm Full scale CM length m
L ft Full scale through-road length m
L fh Full scale heading length m
Lf Fan position with respect to heading entrance m
L mcm Scaled model CM length m
L mh Scaled model heading length m
L mt Scaled model through-road length m
M Mach number
P Pressure Pa
Q air Volume of air m3
Q fs Full scale scrubber exit volumetric flow rate m3/s
Q fj Full scale scrubber exit volumetric flow rate m3/s
Q ft Full scale jet fan volumetric flow rate m3/s
Q ms Scaled model scrubber exit volumetric flow rate m3/s
Q mt Scaled model through-road volumetric flow rate m3/s
Q mj Scaled model jet fan volumetric flow rate m3/s
qd Average dust concentration per shift mg/m3
q Heat flow kJ/m
R Hydraulic radius m
Re Reynolds number
R fj Full scale jet fan penetration m
R mj Scaled model jet fan penetration m
S Mx , S My, S Mz Body forces in directions x, y, z N
td Dust sampling time min
T Absolute temperature K
t Time sec
Tw Wall temperature K
v
SYMBOLS
Greek Symbols
ρ Density kg/m3
μ Viscosity due to linear deformation Ns/m2
λ Viscosity due to volumetric deformation Ns/m2
Φ Dissipation function or deformation heating W/m
υ Kinematic viscosity m2/s
γ Specific weight N/m3
τ Wall shear N/m2
vi
ACKNOWLEDGEMENTS
Acknowledgements
I would like to thank God and the following people and organisations for their support
and contributions to this dissertation:
Dr. Jaco Dirker and Prof. NDL Burger, for leadership and guidance
AEROTHEM, for proving introductory training in Star CCM+
The following persons are also thanked for providing moral support and encouragement:
Florence, my wife
Jones Kharika
Mathias Kyembe
vii
TABLE OF CONTENTS
TABLE OF CONTENTS
ABSTRACT..................................................................................................................... i
ABBREVIATIONS ....................................................................................................... iii
SYMBOLS..................................................................................................................... iv
Acknowledgements ..................................................................................................... vii
LIST OF FIGURES ..................................................................................................... xiii
LIST OF TABLES...................................................................................................... xvii
1. INTRODUCTON AND LITERATURE SURVEY ................................................... 1
1.1 INTRODUCTION .................................................................................................. 1
1.2.1. Mining Activities in Underground Coal Mines .............................................. 4
1.2.1.1. Mining Techniques ......................................................................................... 4
1.2.1.1.1. Bord and Pillar Mining Technique ............................................................. 4
1.2.1.1.2. Long wall Mining Technique...................................................................... 6
1.2.1.1.3. Short wall Mining Technique ..................................................................... 7
1.2.1.1.4. Rib Pillar Extraction Method ...................................................................... 8
1.2.1.1.5. Sub-level Extraction Methods..................................................................... 8
1.2.2. Mining Equipment .......................................................................................... 8
1.2.3. Health and Safety Hazards............................................................................ 10
1.2.3.1. Coal Dust ...................................................................................................... 10
1.2.3.1.1. Description of Coal Dust .......................................................................... 10
1.2.3.1.2. Sources of Coal Dust ................................................................................ 11
1.2.3.1.4. Degradation of Working Environment by Coal Dust ............................... 14
1.2.3.1.5. Coal Dust as Cause of Dust Explosion ..................................................... 14
1.2.3.2. Methane......................................................................................................... 15
1.2.3.2.1. Description of Methane............................................................................. 15
1.2.3.2.2. Sources of Methane................................................................................... 16
1.2.3.2.3. Methane as a Hazard to Miners ................................................................ 17
1.2.4. Legal Requirements and Standards............................................................... 18
1.2.5. Airborne Dust and Methane Control in Coalmines ...................................... 18
1.2.5.1. Ventilation Systems for Development Headings.......................................... 19
viii
TABLE OF CONTENTS
ix
TABLE OF CONTENTS
x
TABLE OF CONTENTS
xi
TABLE OF CONTENTS
xii
TABLE OF CONTENTS
xiii
LIST OF FIGURES
LIST OF FIGURES
Figure 1.1: Perspective view of bord and pillar mining method [3] ............................. 5
Figure 1.2: Perspective view of long wall mining method [3]. ..................................... 7
Figure 1.3: Perspective and closed-up views of a long wall mining system [7] ........... 9
Figure 1.4: A Joy 14 CM Series Continuous Miner [7] .............................................. 10
Figure 1.5: Plan view showing the position of the fan in the coalmine ...................... 20
Figure 1.6: Exhaust duct ventilation system [11]........................................................ 22
Figure 1.7: Forcing duct ventilation system [11] ........................................................ 23
Figure 1.8: Forcing ventilation system with overlap [11] ........................................... 24
Figure 1.9: Exhaust System With Overlap [11] .......................................................... 25
Figure 1.10: A sample converging residual plot [20].................................................... 37
xiii
LIST OF FIGURES
Figure 4.1: Marked positions for tripod stand for the camera................................... 100
Figure 4.2: Block diagram of the generalized measuring system ............................. 103
Figure 4.3: Block diagram of the specific measuring system for unsteady state
experiment............................................................................................... 104
Figure 4.4: Isometric View and a Section of Retro Sensor. ...................................... 105
Figure 4.5: Smoker used for transient state experiment ............................................ 107
Figure 4.6: A Graph of Resistance against Time with Laboratory Lights on and off108
xiv
LIST OF FIGURES
Figure 4.7: Schematic Diagram of Regulator Transformer, Adaptor and LED Lamps
Circuit. .................................................................................................... 109
Figure 4.8: Graph of deviation from average reading against sensor readings for
sensor 2 showing best line of fit, its R-squared value and equation. ...... 111
Figure 4.9: Top View of Points E1, E2 and E3 on MG-2 ......................................... 118
Figure 4.10: Side View of Points E1, E2 and E3 on MG-2......................................... 118
Figure 4.11: Vector diagrams developed from video images for mine geometry, MG-3,
at levels 1, 2 and 3, as seen from the top. ............................................... 120
Figure 4.12: Vector diagrams developed from video images for mine geometry, MG-3,
at sections 1, 2, 3, 4, 5 and 6, as seen from the side. .............................. 121
Figure 4.13: Graph showing the concentration of smoke against time around the
sensors in the scaled model..................................................................... 122
Figure 4.14: Graph showing the concentration of smoke against time around the
sensors in the scaled model after removing malfunctioning sensors...... 123
Figure 5.1: Top view and side view showing flow direction at all points before the
final analysis was done. .......................................................................... 128
Figure 5.2: Plan view experiment and CFD results................................................... 129
Figure 5.3: Front view experiment and CFD results. ................................................ 129
Figure 5.4: End view experiment and CFD results. .................................................. 130
Figure 5.5: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 2. .................................................. 132
Figure 5.6: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 3. .................................................. 133
Figure 5.7: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 4. .................................................. 133
Figure 5.8: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 6. .................................................. 134
xv
LIST OF FIGURES
Figure 6.1: Plan View of the mine showing fan position .......................................... 137
Figure 6.2: Graph of mass fraction of air against time showing the time taken to
extracted pollutants at four different fan positions, i.e.; 0 mm, 80 mm, 500
mm and 1000 mm. .................................................................................. 139
Figure 6.3: Graph of mass fraction of air against time showing that the pollutants were
extracted in 237 seconds at fan position 60 mm. .................................... 140
Figure 6.4: Graph of mass fraction of air against time showing that the pollutants were
extracted in 237 seconds at fan position 100 mm. .................................. 140
Figure 6.5: Graph of mass fraction of air against time showing that the pollutants were
extracted in 192 seconds at fan position 70 mm. .................................... 141
Figure 6.6: Graph of mass fraction of air against time showing that the pollutants were
extracted in 219 seconds at fan position 90 mm. .................................... 142
Figure 6.7: Graph of mass fraction of air against time showing that the pollutants were
extracted in 172 seconds at fan position 75 mm. .................................... 143
Figure 6.8: A summary of time taken to clear the heading at different fan positions
along the length of the heading............................................................... 144
Figure 6.9: Air patterns at three different fan position to illustrate the difference in
time to completely clear out the test section........................................... 146
Figure 6.10: A graph for determining optimum fan position for heading lengths ranging
between 467 mm to 2333 mm. The figure the brackets represent the
equivalent dimensions or measurements in the actual mine................... 148
Figure 6.11: A tool for determining optimum fan position with points x and y. ........ 150
xvi
LIST OF TABLES
LIST OF TABLES
Table 1.1: Methane Produced by Various Ranks of Coal (Source: Curl, 1978)............ 16
Table 1.2: Boundary conditions for compressible viscous flow .................................... 34
Table 2.1: The actual and scaled down dimensions and quantities of coalmine and
equipment...................................................................................................... 47
Table 2.2: Numerical Location of Sensors in the Test Section...................................... 60
Table 3.1: Nominal Loss Coefficients K (Turbulent Flow) [21] ................................... 72
Table 3.2: Reference values that were employed to mesh the model ............................ 89
Table 3.3: Number of Cells for MG 3 for the Three Mesh Models ............................... 89
Table 3.4: Physics Model of Continuum for Steady State Condition ............................ 90
Table 3.5: Physics Model of Continuum for Unsteady State Condition........................ 93
Table 3.6: Boundary Values........................................................................................... 93
Table 4.1: Average of measurements of all sensors taken at different input voltages. 110
Table 4.2: Ohms readings for sensor 2 at different voltage settings, average readings for
all sensors and deviations from the average. .......................................... 111
Table 4.3: Adjustment of measured values for sensor 2 .............................................. 112
Table 4.4: Best line of fit equations ............................................................................. 115
xvii
LIST OF TABLES
Table 6.1: A summary of time taken to clear the heading at nine fan positions...... 144
Table 6.2: Optimum fan position for five different heading lengths and two different
heading widths. ....................................................................................... 147
xviii
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
1.1 INTRODUCTION
Depending on its geological formation, coal can either be extracted from opencast or
underground mining. The ratio between opencast and underground mining in South
African coal industry in 2002 was 48 % and 52% respectively [1].
Like any other mining operations, the extraction of coal, especially underground coal
mining, is associated with a number of hazards. The common problems encountered in
underground coal mining are Airborne Respirable Dust (ARD) and mine gas2.
As coal rocks are brittle, their mining in any form gives rise to a wide range of breakage
products, of which a more or less constant proportion is fine material. Where the
breakage of sandstone or shale, rather than coal takes place, a higher proportion of fine
dust is produced. ARD concentration in the mine is high enough that if not well
controlled it puts the health and safety of the workers and performance of mining
equipment at risk. The dust hazard can be considered from three aspects: explosions and
fire hazards, health risk and nuisance value.
2
Mine gas is also referred to as methane. This is because 80% to 96% by volume of mine gas is methane.
1
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
In the underground coalmine, methane explosions can occur in any place as long as the
conditions are favourable. However, most of them occur at the working faces where
methane emission is the largest. These areas are the gob3, working faces at the
development entries, gob-side tail entry T-junction, near the cutting drum of the
Continuous Miner (CM] or shearer, and in the roof fall cavities.
To prevent mine workers from dust and methane hazards, it must be ensured that the dust
and methane concentration levels do not exceed the recommended standards. In South
Africa, the dust concentration level in the underground coalmine should not exceed 2.0
mg/m³ at operator’s position and methane level should not exceed the maximum
concentration of 0.5% per volume [10].
Ventilation is a popular technique that is used to control dust and mine gas in
underground coalmine. In this method air is utilized to dilute dust and mine gas to
allowable concentration or blow the dust away from mine worker. But, since the heading
is a blind-end penetration, air being blown into the heading is restricted and the air also
re-circulates within the heading making the exercise very difficult.
3
Gob is waste material produced in coal mining, consisting of clay, shale, etc
2
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
One of the problems currently being faced by coal mining operations is the control of
dust and methane in the heading and Last Through Road [LTR]. It is therefore important
to find an efficient ventilation technique that can be applied to control dust and methane
in these areas in the coalmine, to the recommended standards. It is conceivable that by
changing airflow pattern in the mine dust and methane dilution capacity will also change.
Unfortunately, carrying out a research activity like this in the actual mine is almost
impossible. This is because production schedules are very tight, environmental conditions
hush and experimental costs high. Therefore this study seeks to investigate comparatively
inexpensive Computational Fluid Dynamics4 (CFD) modelling technique which can
reliably be utilized in solving ARD and mine gas concentration problems in underground
coalmine.
In section 1.2, some background information is presented to acquaint the reader with
underground coal mining activities and problems that arise from ARD and mine gas.
Brief information on CFD and optimization is presented as well. Problem statement and
method that was utilized in this study is presented chapter 2. Construction of CFD models
is covered in chapter 3. Then practical experiments which were done using a scaled down
coalmine model with the purpose of verifying the CFD model is presented in Chapter 4.
Chapter 5 analyses and compares CFD and experimental results. Further numerical
analysis with an objective devising a method for determining optimum position of jet fan
is covered in chapter 6. The dissertation finishes with conclusions and description of
future work and related topics that arose from this study.
4
CFD is the analysis of systems involving fluid flow, heat transfer and associated phenomena such as
chemical reactions by means of computer-based simulation [20].
3
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
In the early days coal was obtained by digging it from where it appeared on the surface.
Later the first small mines were sunk known as bell pits. As mining techniques constantly
improved, working into deeper seams became feasible. The two major types of mines are
underground mines and surface or opencast mines. This dissertation is dealing with
underground coalmines.
In underground coalmine different extraction methods are used depending upon the
geological formations of the coal and other factors. The two major extraction methods are
Bord and pillar (also referred to Room and Pillar) and Long wall. The two methods are
sometimes combined to come up the third method known as Short wall.
Bord and pillar mining is the most common type of underground coal mining method in
the world, see Figure 1.1. [3]. Cutting operations are performed using a Continuous
Miner (CM) that cuts out extended cuts into the mining face. To complete a cutting
operation the CM makes four cuts, each 17.5 m long and 3.5 m wide into the coal seam.
These cuts are known as box cuts. The final depth of the heading might be up to 35 m,
the width up to 7 m and the height depends on the thickness of the coal seam.
As the rooms are cut, the CM simultaneously loads the coal onto a shuttle or ram car
where it will eventually be loaded on a conveyor belt that will move it to the surface.
Each room alternates with a pillar of greater width for support. Using this method
4
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Figure 1.1: Perspective view of bord and pillar mining method [3]
Bord and pillar mining has been, and still is, most widely practiced method of
underground coal mining in South Africa owing to its inherent safety, low capital
investment, and low operating cost [4]. The Twisdraai coalmine, part of Sasol Mining’s
Secunda Collieries5 complex is one example of a board and pillar operation. At these
mines rooms, also referred to as headings during cutting process, are accomplished by
four cuts, as already explained above.
5
Coalmine with its buildings and equipment
5
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
In opening a mine on the bord and pillar system, gangways, entries, headings or galleries,
as they are variously called, are first run forward on the face and end slips of the coal
seam. The entries constitute the main avenues of the mine; they are usually driven much
narrower than the rooms so as to make them safer, as well as to add strength to the pillars.
In the headings breakthroughs, also referred to as Last Through Road (LTR), need to be
made from one heading to another at stated intervals for the passage of the ventilating
current of air and other purposes.
With long wall mining methods the principle is to extract all of the coal over the width of
the panel face in successive cuts, see figure 1.2. After the coal is removed, it drops onto a
chain conveyor, which moves it to a second conveyor that will eventually take the coal to
the surface. Temporary hydraulic-powered roof supports hold up the roof as the
extraction process proceeds. This method of mining has proved to be more efficient than
conventional room and pillar mining, with a recovery rate of nearly 75 %, but the
equipment is more expensive than conventional bord and pillar equipment, and cannot be
applied in all geological circumstances. As mining continue, only the roofs for main
tunnels are bolted to avoid ceiling collapse. The rest of the roof is supported by large
steel supports, attached to the mining machine. As the machine moves forward, the roof
supports also advance. The roof behind the support is allowed to fall.
6
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
The short wall method combines many of the features of bord and pillar and long wall.
The area to be mined is developed by bord and pillar equipment, but the blocked-out
pillars are larger. The pillars are extracted by taking wide slices off a relatively short
wall, the roof being controlled by heavy-duty powered supports of the type commonly
used on long walls.
The primary difference between long wall and short wall methods is that short walls are
generally 46 to 61 metres wide, while the long walls are 106 to 182 metres wide [3].
Long wall mining can be practiced as an advancing or a retreating system (although only
the latter currently is being used in South Africa) while short wall mining is usually only
practiced on the retreat. In 2002, South Africa had three mines, namely, Malta No. 2,
Amot No. 8 shaft and New Denmark, Okhozini Shaft, extracting coal by short wall
method [1].
7
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
A major advantage of this method is that the same equipment, with the exception of the
powered supports, can be used in both the development and pillar extraction phases of
operation. And the capital cost of the supports is lower than for a long wall [6].
Rib pillar extraction refers to a series of methods that can be regarded as a combination of
pillar extraction and short wall methods.
The term ‘rib pillar’ was coined in South Africa to describe a series of methods that are
based on the extraction of a rib of coal between development roads and the goaf6, with a
solid block of coal providing the major means of support in the workings.
Sub-level extraction in coal mining usually is applied only in coal seams where the nature
of the seam excludes the practical application of other coal mining methods.
The method basically consists of driving a series of sub-levels commencing at the top of
the ore body. A starting vertical slot is cut and then a series of ring patterns are drilled
and blasted, the broken coal being drawn off after each blast. In South Africa there are
very limited deposits of coal that would be suitable for this application.
6
The empty area left by the extraction, which, however, is soon filled up by falls of stone from the strata
overlying the bed. See figure 1.3.
8
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
o Continuous miners and Shearers: These are the major cutting machines.
Continuous Miners are usually used in room and pillar-mining method while
Shearers are popular in long wall.
o For transporting coal out of the mine, loaders, shuttle cars, continuous haulages
and conveyors are usually employed.
o For health and safety purposes, roof supports, dust collectors, ventilation
equipment (e.g. jet funs) are used among other equipment.
Components of the long wall mining system and the continuous miner are shown in
figures 1.3 and 1.4 respectively.
Figure 1.3: Perspective and closed-up views of a long wall mining system [7]
9
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Airborne Respirable Dust (ARD) and methane are the two common problems
encountered in underground coalmine. In the quantitative risk assessment for thick-seam /
long wall operation [1], using field data, dust was identified as both an explosion and a
health risk in the rankings of 2 and 1 respectively, see Appendix A.
The fine coal and rock particles produced in underground mining are called dust. The
sizes of the dust particles are denoted by their average diameter in micrometre,
abbreviated µm (1 µm = 10-6 mm). The dust that is suspended in and moves with the air
ventilation flow is the airborne dust and that which settles down on the floor is the settled
10
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
dust. Dust concentration is the amount of airborne dust contained in one unit of air
volume. It is generally represented by the total weight in milligrams per cubic metre
(mg/m³) or sometimes by number of dust particles in one cubic centimetres of air volume.
In this document the former is used.
The amount of airborne dust made per tonne of raw coal produced, Ad, can be determined
by:
q d Qair T
Ad = (1.1)
As
Where qd is the average dust concentration per shift in mg/m³, Q air is the air volumetric
flow rate in m³/min at the location where the dust concentration is measured; T is the dust
sampling time in minutes, and As is the shift production in tonnes.
The amount of airborne dust and its duration in the airflow depend not only on the sizes,
shapes, and weight of the dust particles, but also on the air velocity and humidity. As
these factors change, the airborne dust may be converted to the settled dust or vice versa
[1].
Mechanized coal and stone operations, more concentrated workings, wide spread use of
powered supports and thus more caving of wastes with a more rapid advance of coal face,
have all contributed to the increased production of dust in recent years [2].
Most of the dust produced is formed and dispersed at the coalfaces, and at various places
in the intake airways, which thus contaminate the air supplying the coalfaces.
11
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
o Mechanical breakage and disintegration of the coal and adjacent strata during
mining operations, and in some cases dispersion during mining operations of
dust already present in the slips and bedding planes.
Since coal rocks are brittle, their mining in any form gives rise to a wide range
of breakage products, of which a more or less constant proportion is dust.
The coals are further broken down and more dust is produced during
transportation by conveyors and loaders, especially at the transfer points.
In addition to the basic sources of dust formation and dispersion there is also the problem
of disturbing and re-dispersing dust that has settled out of the ventilation current. Dust
can be re-dispersed by high air velocities or by the movement of people and machinery.
12
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
If miners are exposed to ARD for a long time, the inhaled coal dust could cause health
hazards ranging from minor bronchial disorders to pneumoconiosis or silicosis. The latter
is the most widely recognized occupational disease in coal mining and is commonly
referred to as Coal Workers’ Pneumoconiosis (CWP) or black lung. The concentration of
the CWP can be attributed to the following factors:
o Dust Size: The airborne dust contains various sizes of dust particles. Once the
airborne dust is inhaled by the human being, due to their own weight and air
impact, the larger particles will settle on the membranes of the bronchial tubes
before they reach the lung. On the other hand, the fine particles, that is, those
with diameter less than 5 µm, will be able to reach the lungs. These fine
particles account for a large proportion of the airborne dust, sometimes up to
80% [1]. The fine particles that can penetrate into the nonciliated space in the
lung and lead to pneumoconiosis are the respirable dust.
o Free Silica: The amount of free silica in the airborne dust is directly related to
the rate of development and severity of the pneumoconiosis. But only a small
amount of free silica is associated with the airborne dust in the coalmines.
o Dust Concentration: The amount of respirable dust that enters the lung is
proportional to the dust concentration and the duration of contact with the
airborne dust. Therefore if effective methods are employed for controlling the
dust generation and/or dust flows, the dust concentration can be greatly
reduced. Under this condition, the amount of respirable dust inhaled will be
too small to cause pneumoconiosis even after long-term exposure.
13
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Airborne dust particles of all sizes may cause irritation of the eyes, ears, nose and throat,
and also result in skin irritation where the temperature and humidity are high. Small
electrical relays, switches, hydraulic circuitry, machine and motor bearings can all suffer
from abrasive dust and cause delays, with a possible loss of production. If the air at the
face area is filled with airborne dust, it is very difficult to make the visual inspection
required constantly for operating the machines and roof conditions. This may lead to
accidents.
When the concentration of the airborne dust with particle sizes ranging from 25 - 100 µm
reaches certain limits, dust explosion will occur provided the following conditions
prevail:
o The coal dust is more liable to explode as the amount of volatile matter in the
coal increases. But this cannot be used as the only criterion because the
composition of coal is rather complicated.
o All airborne dust particles smaller than 100 µm contribute to dust explosion.
But the most dangerous range is from 25 to 75 µm.
o The presence of methane will lower the limit of dust concentration for
initiating dust explosion.
14
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Therefore coal dust constitutes a large threat to the health and safety of coal miners. It
must be efficiently and effectively controlled.
1.2.3.2. Methane
In general, mine gas refers to all the hazardous gases in mines. The most frequently
encountered hazardous gases in underground coalmines are methane (CH4), carbon
monoxide (CO), sulphur dioxide (SO2), hydrogen sulphide (H2S), nitrogen dioxide (NO2)
and hydrogen (H2).
Methane is the major component of the hazardous gases in the underground coalmines
occupying approximately 80 – 96 percent by volume. Methane is a colourless, odourless,
flammable gas. Its diffusivity7 is about 1.6 times that of air. Since it has a low specific
gravity8 (0.554), methane is easily accumulated near the roof of the roadway and working
faces in the mine.
7
Diffusivity is classically defined as the mass of solute transferred per unit area per unit time under unit
concentration gradient.
8
Specific density is the ratio of the density of a given solid or fluid substance to the density of water at a
specific temperature and pressure. It is unit less.
15
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
During the coalification process, the coal forming materials, which consist mainly of
plants, undergo a series of physical and chemical reactions and produce a large volume of
methane. After that, the coal is subjected to metamorphism and undergoes changes in
chemical composition and structure under high pressures and temperatures.
Metamorphism process also produces methane. In general the amount of methane
produced is proportional to the ranking of metamorphism, see table 1.1.
Table 1.1: Methane Produced by Various Ranks of Coal (Source: Curl, 1978)
Low High
Volatility Volatility Semi
Rank Lignite Bituminous Bituminous Anthracite Anthracite
Methane
(m³/tonne) 68 230 270 -290 330 400
Most of the methane produced during coalification and metamorphism escapes to the
atmosphere through fissures in the strata. A small amount stays in the fissures and still
16
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
another small amount remains in the coal. This is the methane that is commonly referred
to mine gas.
Under a certain temperature, methane will be ignited when its concentration is between 5
and 15 percent. When methane concentrations are 5%, it forms a bluish stable
combustion layer around the flame without initiating explosion. When the methane
content is larger than 15%, there is insufficient amount of oxygen to promote explosion
[3].
The study conducted by John Frint in 1990, reviewed that between 1891 and 1990, a total
of 155 explosions occurred in South African collieries that resulted to 799 fatalities [9].
Details are shown in Appendix B.
Based on the factors contributing to methane explosion, most of the preventive methods
of methane explosions are based on reduction of methane accumulation and elimination
of high temperature heat sources. Flint identified the heat sources for mine gas explosion
and suggested their prevention methods. This study dwells on the control of methane and
dust concentration by ventilation method.
17
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
One of the most important features in the control of environmental hazard is the
establishment of standards by which the quality of relevant aspect of the environment
may be judged objectively.
In the coal mining industry worldwide there are also different legal requirements and
standards for ARD and methane concentration for different countries. The United States
Federal Coal Mine Health and Safety Act of 1969, limits personal exposure to respirable
dust to 2.0 mg/m³ and the concentration of methane 1%. A directive of the South Africa
Department of Minerals and Energy (DME, 1997) required the dust concentration level to
be reduced below 5 mg/m³ at the operator’s position [10]. This figure is now being
reduced further to 2 mg/m³. And methane concentration in South African coalmines
should be kept below 0.5%.
Coal Dust Control – Dust control in the coalmines can be divided into three major
categories: ventilation, water, and dust collection. Since this research is mainly concerned
with ventilation method, more details are given on this area while general information is
covered on the rest.
18
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Methane Gas Control - There are three requirements for methane explosion: minimum
concentration of methane, minimum concentration of oxygen and a suitable heat source,
as mentioned before.
Based on the factors contributing to methane explosion, most of the preventive methods
of methane explosions are based on reduction of methane accumulation and elimination
of high temperature heat sources. This study dwells on control of methane concentration
by ventilation method.
Methane Control – The primary function of ventilation is to dilute and remove mine gas.
The amount of air required for methane dilution is greater than that for dust reduction.
For example, in the USA the fresh air required at the long wall face is from 510 to 1417
m³/min. In methane rich seams, it is not unusual to have the fresh air intake at the rate of
1840 m³/min [3]. Ventilation also reduces the temperature in the mine since high-
temperature contributes to methane explosion.
Dust Control - Ventilation air reduces dust through both dilution and in some cases,
displacement. The dilution mechanism operates when additional air serves to reduce the
dust concentration by diluting the dust cloud surrounding the workers. The displacement
mechanism operates when workers are upwind of the dust source and the air velocity is
high enough to reliably keep the dust downwind.
The minimum quantities of air are prescribed by the Mines and Works Regulations,
which require that, throughout the 24 hours, the face should be provided with
0.001m³/sec of air per 25 multiplied by the mass of coal mined in tonnes per shift, and
further that the velocity of air over the working height shall be not less than 0.25 m/sec
[4]. The velocity of the air current at which dust pick-up may start to be troublesome is
19
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
2.5 – 3.0 m/s9. Therefore it should be ensured that the air speed must not reach dust
picking levels. This conflicts with the necessity of removing methane [4].
Jet fans are predominantly used for longitudinal ventilation and are positioned at heading
entrance near to the wall opposite the CM as shown in figure 1.5. The standard position
for the jet fan is 1.2 m from the heading entrance, 1 m from the side wall and 0.7 m from
the floor. Modern jet fan have variable pitch and speed control as well as reversible flow
direction.
Figure 1.5: Plan view showing the position of the fan in the coalmine
9
This figure however depends on the inherent moisture in the coal, particle size and method of working.
20
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Due to the heading being blind air from the fan end up circulating and re-circulating
resulting into inefficient control of ARD and mine gas. To overcome the re-circulation
problem some auxiliary equipment like ducts are sometimes used.
Ventilation by means of duct can be exhaust, forcing or overlap; these are described in
details in the following sections:
Exhaust System
Exhaust ventilation system consists of main exhaust duct, to which a duct fan is attached
on its suction side in the heading as shown in figure 1.6. With exhaust system relatively
clean air passes the miners and the extracted dust is fed via the fan into the return air
stream. There are, however, two disadvantages. Firstly, if firedamp is a problem, pockets
of mine gas could accumulate at the face of the heading, together with possible roof
layering because of the low velocity and limited scouring action of the airflow.
Secondly, the dust exhausted from the face of the heading may contaminate the
ventilation used to supply other working places. The second problem may be overcome
by using dust collector to take out a large part of the airborne dust before the air is used
again [2]. However, if gas emission is a problem, then an overlap system should be
considered.
21
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
An exhaust ventilation system can be equipped with extra ventilation devices, such as: air
curtain on a CM, jet fan in working face zone and air cooler at a certain distance from the
face zone [11].
Forcing System
A forcing ventilation system consists of the main duct line, to which a duct fan is
attached on its forcing side in the heading with through air shown in figure 1.7. Forcing
ventilation is characterized by blowing a greater quantity of airflow rate to the heading
face than exhaust ventilation. This makes it the best system for dealing with firedamp
[11]. The forcing system, however, has the disadvantage of exposing the miners to
airborne dust of maximum concentration on its return journey along the roadway [2]
22
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Forcing ventilation system can be equipped with extra ventilation devices, such as: air
cooler with an auxiliary fan located in the main forcing duct and/or a cooler in an
auxiliary duct located in the working face zone [11].
The forcing system with overlap, shown in figure 1.8, is the one recommended where
possible [2]. The forcing section is the main part of the ventilation system and the short
exhaust section is the secondary part for dust control, usually incorporating some form of
dust collector. The forcing section gives a good scouring action at the face of the heading.
It is normally necessary to use a diffuser at the delivery end of the forcing ducting to
reduce the delivery velocity and dust pick-up. When it is necessary to cool air, a cooler is
placed in the forcing duct [11].
23
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
As with forcing system with overlap, the exhaust system with overlap, shown in figure
1.9 combines the advantages of having a good scouring action from the forcing section,
with the removal of the high dust concentration to bypass the workers. This system has
the following solutions [11]:
o Short forcing duct with a vortex duct and duct hopper in the face zone in case of
methane hazard.
o Short forcing duct with a vortex duct, a duct hopper in the face zone and a jet fan
in the heading in case of extremely high methane hazard.
24
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
A part from individual problems of fan and duct system mentioned under specific
systems the general disadvantage of this system is that the equipment has to be adjusted
as the heading depth progresses. Considering the harsh working environment in the mine,
workers find it difficult to move the equipment around as a result this system is not
applied in most of the mines.
Ventilation is just one common method of controlling ARD and mine gas but there are
also other methods that are used in addition to ventilation. This section describes some of
the other popular methods.
25
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Water is also used even to control airborne dust like in the wet scrubber and water
curtains.
Dust collectors are used to remove dust from air stream. Scrubber is the most popular
dust extractor in the underground coalmines. Recently wet scrubber has proved to be
more efficient as far as dust extraction is concerned.
1.2.5.2.1.3. Surfactants
Surfactants are chemicals that are added to spray water in order to increase its
effectiveness in terms of dust suppression. Surfactants help to bond the dust particles
together so that they are not easily become airborne in case of high air velocities or mine
explosion.
26
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Good cutting techniques by using appropriate cutting machine for a specific coalmine
help to reduce the amount of dust production. For more information see references
[1][2][3] and [6].
Usually, methane accumulates to high concentration in the gob. In the USA the gobs are
ventilated to prevent methane accumulation and to reduce temperature. In most other
countries the gobs are tightly sealed such that it completely cuts off any fresh air flowing
into the gob or prevents high-concentration methane flowing out of the gob.
If the coal seam has high methane content methane drainage is considered. Methane
drainage involves drilling boreholes into the solid coal, the roof and sometimes the floor.
The withdrawn methane is used as fuel or as a raw material for chemical by-products.
The other method of preventing the accumulation of methane in the mine is water
infusion. This involves drilling in-seam horizontal holes into the solid coal ahead of
mining. High-pressure water is injected into the boreholes and moves away in a
cylindrical waterfront. As the water moves away from the borehole, the methane is also
driven away.
Since health and safety risks caused by ARD and methane where identified in coal
mining numerous research work have been conducted in order to mitigate or eliminate
their effects. In addition to the work described in the preceding sections this section gives
27
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
a summary of other works in the area of mining ventilation that is more relevant to this
particular research.
o The benefit of using a simulation tool such as VUMA-Network is that the results
can be viewed graphically.
o The effect of various changes to the existing ventilation network can be easily
assessed.
o Real benefits can be derived in areas of airflow distribution and fan power
requirements by using VUMA-Network.
28
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
The extensive full-scaled simulation studies in the ventilation tunnel at the CSIR resulted
in a new Kloppersbos Spray System and the development of new dust-control systems.
They are the half-curtain, the retrofitted hood system, double scrubber system and the
integrated system [14].
According to Bell, in more than 80% of the new dust-control systems were being
operated in South Africa mines in 2002. The half-curtain dust-dust control system was
the most successful of them all. Test carried out on the half-curtain system in the
Kloppersbos ventilation simulation tunnel showed that effectively dilutes the methane
released at the face at a rate of 600 L/min (Van Zyl et al., 1999).
o Hollow-cone spray nozzles with inlet and outlet diameters of 1.6 and 2.0 mm
respectively.
29
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
As already defined in footnote 4, CFD stands for Computational Fluid Dynamics and it
deals with the entire study field of fluid mechanics using computational method.
In this work, the author would like to investigate if CFD can be used to study air flow in a
scaled-down underground coalmine. It is for this reason that the basic differential
equations on which all CFD packages are built introduced.
The fundamental equations for compressible viscous flow of a fluid are based on the
three laws of conservation for a physical system [15]:
The three unknowns that can be obtained simultaneously from these three basic equations
are the velocity V, the thermodynamic pressure p, and the absolute temperature T.
Thermodynamic pressure and absolute temperature considered as independent
thermodynamic variables. However, the final forms of the conservation equations also
contain four other thermodynamic variables: the density ρ, the enthalpy h (or the internal
energy ei), and the two-transport properties µ (viscosity) and k (conduction).
In order to specify a particular problem completely, the conditions (of various types) for
V, p and T must be known at every point of the boundary of the flow regime. The
30
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Please note that all the three basic laws are formulated for particles (systems) of fixed
identity. Thus, in the Eulerian system appropriate to fluid flow all the three laws utilize
the particle derivative
D δ
= + (V • ∇) (1.2)
Dt δt
The following partial differential equations were derived for general control volumes,
expressed in Cartesian coordinates (fn):
Dρ
+ ρ. ∇ • V = 0 (1.3)
Dt
with
δu δv δw
∇• V = + +
δx δy δz
31
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
The Navier – Strokes equations express proportionality between applied force and the
resulting acceleration of a particle of mass. This is the reason it is also commonly known
as Newton’s second law.
Du δp δ δu δ δu δv δ δw δu
ρ = S Mx - + {2μ +λ ∇ • V}+ [μ{ + }] + [μ{ + }]
Dt δx δx δx δy δy δx δz δx δz
(1.4)
Dv δp δ δv δu δ δv δ δv δw
ρ = S My - + [μ{ + }+ {2μ + λ ∇ • V }+ [μ{ + }]
Dt δy δx δx δy δy δy δz δz δy
(1.5)
Dw δp δ δw δu δ δv δw δ δv
ρ = S Mz - + [μ{ + }]+ [μ{ + }]+ {2μ + λ ∇ • V }
Dt δz δx δx δz δy δz δy δz δz
(1.6)
32
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
With the assumption that Fourier’s law governs the heat transfer to the element volume10,
the energy takes the final form of:
Dh Dp
ρ = + div (k∇T) + Φ
Dt Dt
(1.7)
The basic equations given in this section are used to derive CFD codes.
Any CFD model is as good as the assumptions it is based on. One can have a sufficiently
converged solution, but if the boundary conditions do not resemble the reality, the
solution is useless. Thus, it is very important to apply the correct boundary conditions.
In this research were the fluid in question is incompressible and viscous (air), the
boundary conditions are given in table 1.2.
10
Fourier’s law: q = -k∇T where: q = the vector heat flow per unit area in W/m2; k = conduction of the
continuum in W/mK; T = absolute temperature in K/m.
11
The dissipation function Φ involves the viscous stresses. It is always positive definite [15].
33
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Solving complex fluid flow problems using the differential equations requires a
numerical approach. A numerical solution of a differential equation consists of a set of
numbers from which the distribution of the dependent variables can be constructed. This
is different from the analytical solution that describes the continuous values throughout
the domain thus an infinite amount of values of the dependent variables.
In numerical method the governing Partial Differential Equations (PDE) are replaced
with simple algebraic equations, which can be solved with relative ease. This is achieved
by replacing the continuous information contained in the exact solution of the differential
equation with discrete values (of dependent variables) at a finite number of given points
in the domain. This is referred to as discretisation.
most common are the finite difference, finite element and spectral methods. In this
dissertation the Finite Volume Method (FVM) formulation, which falls under finite
difference method is used. This is because the resulting statements of the control volume
integration express the exact conservation of relevant properties of each finite size cell
and is central to the most well established CFD codes like CFX/ANSYS, FLUENT, EFD
Lab, PHOENICS and STAR-CD. Since FVM is used in this research as a solving
method, it is the only one that has been described in this dissertation. For the sake of the
readers who are interested in the other methods, refer to references [15-19].
o Integration of the governing equations of fluids flow over all the finite control
volumes of the domain.
Any CFD flow problem starts with the preparation of geometry using CAD software, as
well as dividing the geometry into cells, called the mesh or grid before any modelling is
done.
Most commercial CFD packages use their own pre-processor to generate both the
geometry and mesh.
35
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
The Finite Volume Method (FVM is one of the most versatile discretisation techniques
used in CFD. Based on the control volume formulation of analytical fluid dynamics, the
first step in the FVM is to divide the domain into a number of control volumes12 where
the variable of interest is located at the centroid of the control volume. The next step is to
integrate the differential form of the governing equations over each control volume.
Interpolation profiles are then assumed in order to describe the variation of the concerned
variable between cell centroids. The resulting equation is called the discretised or
discretisation equation. The discretised equation expresses the conservation principle for
the variable inside the control volume and its result satisfies the conservation quantities
such as mass, momentum, energy, and species. [19]
The quality of simulation results depends upon convergence and grid independence. A
solution of the algebraic equations that approximate a given PDE is said to be convergent
if the approximate solution approaches the exact solution of the PDE for each value of
the independent variable, as the grid spacing tends to zero.
12
Also referred to as cells or elements
36
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
In Star CCM+ convergence is judged by residual monitor plots. The residual in each cell
represents the degree to which the discretised equation is not completely satisfied. The
residual plot is a monitor plot that is automatically created from the active residual
monitors on semi-log axes when iterating starts [20]. A sample residual plot is shown in
figure 1.10.
Residual quantity will tend toward a very small number when the solution is converged
and will increase in number when it is diverged.
37
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
Evaluating the base case usually gives the user insight into the problem and can lead the
user to identify suitable parameters or variables that have a marked influence.
Optimization seeks to change design variables of an identified base case in such a way
that the objective function is achieved. In choosing the objective function it must be
ensured that its parameters or variables can be minimized for the optimum solution.
Constraint functions that restrict the achievement of the set objective are also identified.
The more variables to be optimized, the more perturbations13 are essential for the
optimizer to predict the next optimum design iteration. Parameterization is can be done
by most of the CFD commercial software.
13
Small changes in terms of optimization process.
38
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
The optimization process begins by iteration of the base case and its perturbations and
moves on until the established objective function is approximated. In other words, the
solution to the preceding iteration is used to do to succeeding iteration. This process is
carried on until the objective function converges to a constrained minimum.
In case the validation using the actual system or process results into high financial,
environmental or any other undesirable implications, other experimental validations can
be considered. In this dissertation, the CFD model was validated with the experimental
results done using scaled down physical model of an underground coalmine in the
aerodynamics lab at the University of Pretoria.
The literature study has reviewed the following concerning ARD and methane problem in
the underground coalmines:
39
CHAPTER 1:INTRODUCTION AND LITERATURE STUDY
o Most of the dust control techniques were developed in the USA for low-seam
long wall mines (seam heights up to 2.5m). The applications of these techniques
are limited to RSA high-seam conditions in which the operating seam height
reaches 4.1m.
40
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
2.1. Introduction
This chapter explains the problem under investigation in details and explains the method
that was used in the study in order to arrive at solutions.
As it has been mentioned in the previous chapter, coal dust and methane gas are a hazard
to coalmine workers and mining equipment.
To ensure safety of coalminers the South African government through the Department of
Energy stipulated that dust concentration level in the underground coalmine should not
exceed 2.0 mg/m³ at operator’s position and methane level should not exceed the
maximum concentration of 0.5% per volume [10]. A number of initiatives have been
developed to mitigate these hazards but none seems to work effectively. Therefore, work
has to continue to ensure that the standards are met or even exceeded for the benefit of
the miners. Methods that are employed to control the hazards have been described in the
literature study of which ventilation is popular.
41
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
a means of carrying meaningful investigations at low cost and less time. One way of
achieving this is by using computer modeling and simulation.
This work was done in order to find out if CFD numerical method could be used to carry
out ventilation related investigations for the underground coalmine.
2.3. Methodology
Numerical solutions can only be relied upon with confidence if they can be validated with
experiments. Therefore, in this investigation experiments were conducted to validate the
CFD results. Validation was achieved by comparing the numerical and the experimental
results. Similarity of the results validates the numerical method. Two types of
experiments and their related CFD simulation were done. They are steady and unsteady
or transient states. The CFD models are discussed in chapter 3 and experimental model in
chapter 4. Comparison between CFD and experimental results are captured in chapter 5.
In order to reduce the cost and time for the exercise a scaled down physical model was
used instead of the full scale model. The scaling down of the full scale model is
described first before experiment methods. It should be understood that the scale model
was not used to represent a full scale mine but rather to validate a CFD model. This can
later be extended to find out if the concept is applicable to a full scale mine later.
It should be noted in advance that scrubber was included in this study but it was difficult
to visualize clearly the flow of air in the steady state experiments therefore it effects are
not shown in steady state results. But the results in unsteady state include the effect of
scrubber.
42
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
The scaled model that was originally designed and built by A.A. Daling [25] was used.
The model was scaled down to 15% and is located in Aerodynamics Laboratory at the
University of Pretoria. This section describes the concept, the scaling down and
modification work done on the model in order to achieve the objectives.
The model shown in figure 2.1 comprises: a through-road, a heading, CM, jet fan and
scrubber.
43
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
No Description No Description
1 Through-road 6 Jet fan
2 Heading 7 Smoke System
3 Continuous Miner 8 Table
4 Scrubber 9 Partitioning mechanism
5 Flexible duct
The model was scaled down to 15% from the dimensions and quantities of actual mine as
described in the next section.
The tapered end of the through-road on the model (please note that the actual through-
road has not tapered end) is fitted with a fan that facilities air velocity in the through-
44
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
road. The tapered end was done for the purposes of creating a well developed flow in the
through-road considering the size and shape of the fan used as compared to the size and
rectangular shape of the through-road cross section.
It was not possible to find a jet fan scaled down to 15% with the same penetration as the
actual fan; therefore the jet fan is represented in the model by a nozzle blowing air from a
compressed air system [25]. The exit diameter of the nozzle that gives a similar
penetration of the jet fan is 10 mm as calculated in 2.4.2.4.[25]
The left half of the coal seam, as seen from heading entrance, was made in such a way
that it can be adjusted forward or backward. See a picture of the scaled model in figure
2.2. An addition mechanism to block the right half was also provided for in the design.
These two mechanisms facilitated modification of model geometry to different shapes as
the case in the actual mine when mining process is in progress.
Figure 2.2: A picture of scaled-down coalmine model showing adjustable part of the
heading representing coal seam.
45
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
The top and right side of the heading wall, as seen from heading entrance, are made of
transparent plastic sheet. This is so to facilitate convenient observation of air flow when
conducting the experiments. In addition, the top part can easily be removed in order to
access the inside of the heading for the purposes of changing the geometry of heading,
fixing sensors, setting the smoke tubes, etc.
2.3.1.2. Scaling
According to T. Prinsloo [25] who did an investigation for optimisation of air flow
patterns in coalmines using the same scale model, the conversion of data from full scale
model to 15%-scale model using physical phenomenon result into impractically high
quantities. For example, using Reynolds number the exit air speed for scrubber was found
to be 152 m/s. Therefore, air flows were scaled using percentage volume flow method.
The percentage volume flow rate is the amount of air that is replaced by new air in
relation to the entire volume of air in the volume of interest. And dimensions for the mine
and CM were scaled down by geometrical method.
Please note that the intension here is not to represent a full scale mine by sing a scale
model, but rather to validate a CFD model which can later be used to represent a full
scale mine.
The following sections describe how the scaled down sizes and quantities, shown in table
2.1, were found as inspired by a full scale mine:
46
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
Table 2.1: The actual and scaled down dimensions and quantities of coalmine and
equipment
The dimensions of the mine geometry and CM were scaled by a scale factor of 1 : 15.
This means that there is a linear relationship between the actual mine and the scale
model.
14
Please take note that the actual width of most mines in South Africa is 7 m but 6.3 m in this research
because the model by A.A. Daling was scaled from this dimension.
15
The length of the scaled model represents a segment of 18 m.
16
A reach is a jet penetration depth.
47
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
Lfh = 15 × Lmh
35 = 15 × Lmh
Lmh = 2.33 m
2.3.1.2.2. Scrubber
The scrubber was scaled by percentage of volumetric flow method. To determine the
flow of the scrubber in the scale model the percentage volumetric flow rate of the scale
model is assumed to be the same as that of the full model. Scrubber fan diameter is scaled
geometrically by a factor of 1:15. The scaling is as follows:
π
Q fs = 35 × × 0.62
4
3
= 9.9 m /s
Q fs
P% =
( L × W × H ) fh
9. 9
=
35 × 6.3 × 5
= 0.009
48
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
Qms
P% =
( L × W × H ) mh
Q ms = 0.0029 m3/s
Scaling down the exit diameter for scrubber the result is 0.04m, therefore the exit
velocity,
Qms
V ms =
Ams
0.0029 × 4
=
π × 0.04 2
= 2.3 m/s
Through-road is ventilated by the main ventilation system and flow is usually 2 m/s.
There is generally no dust and mine gas concentration problems in the through-road since
air flows are sufficient enough to dilute mine gas and dust. Though there is no current
concern with through-road ventilation, it is brought into picture here because the air
flowing in the through-road influences heading flow. Flow in the through-road is scaled
by percentage of volumetric flow method as follows:
Q ft = 2 × 6.3 × 5
= 63 m3/s
49
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
Q ft
P% =
( L × W × H ) ft
63
=
(18 × 6.3 × 5)
= 0.111
The through-road volumetric flow rate for the 15%-scale model is,
Qmt
P% =
( L × W × H ) mt
Q mt = 0.019 m3/s
V mt = 0.0665 m/s
For the jet fan scaling, it is important to take into consideration the fan penetration in
order to arrive at accurate values. In the actual mine which is used as reference for the
scale model, the jet fan with an exit velocity of 45 m/s only penetrates approximately 26
m into the blind heading [25].
50
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
As it has already been explained in 2.4.1, the jet fan in the scaled model was replaced by
a nozzle with exit diameter of 10 mm17, blowing out air from air compression system.
The jet fan properties for the scale model were calculated and are shown in table 2.1. The
standard and scaled down positions of the fan are also indicated in the same table. The
width of test section is half the total distance because it is assumed that the air will need
equal cross section areas into and out of the heading.
π
Q fj = 45 × × 0.2662
4
= 2.5 m3/s
= 0.0041
17
It should be noticed that this figure is not found by geometrical scaling but it was found by comparing jet
penetrations of different exit diameters[25].
51
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
V mj = 9.0 m/s
As air is transparent and cannot be seen with human eyes, a smoke producing device and
a system to conduct the smoke into the test section18 of the model was designed. This
system facilitates the visualisation of air flow. Figure 2.3, shows a schematic
representation of the smoke producing device. It comprises a container with a heating
element inside it, a glycol/water mixture19 reservoir with a submersible pump in side it, a
network of tubes and a fan that blows smoke into the test section.
18
The mined spaced in the heading including through-road is the test section.
19
Glycol/water mixture has commercial names called fog fluid, fog juice or smoke water. It commonly
used by Disk Jockeys to create a foggy environment at discos or parties. One can easily get this fluid in
music equipment shops in RSA.
52
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
No Description No Description
1 Tubes delivering smoke into heading 6 Glycol/water mixture return tube
2 Main smoke tube 7 Glycol/water mixture return tube
3 Heating Element 8 Glycol/water mixture reservoir
4 Smoke extractor 9 Stand
5 Smoke container 10 Pump
Figure 2.3: Smoke System using a pump to spray fluid onto the heating element
53
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
The smoke system works as follows: glycol/water mixture is pumped from the tank into
the container where it is sprayed onto the hot heating element. The burning of the fluid
produces smoke. The smoke is blown into the test section by the fan and the liquid
glycol/water mixture returns to the circulation process. The smoke enters the test section
through the tubes inserted in the holes on the heading floor of the model as indicated in
figure 2.4.
Initially, the heating element in use was a modified domestic aquarium heater with a
rating of 75W. In the course of performing experiments, the pump broke down due to
high temperature of the fluid. As a result the heater burnt out. The heater was then
replaced by a coiled heating element with a rating of 1000W because a replacement that
could fit the pot design was not readily found.
All the walls inside the scale model, excluding the transparent observation panels, were
painted black (using non-gloss paint) to enhance visibility of smoke when capturing
video images.
Despite the scale model being available for the required experiments, it was found that it
was not in working condition. The following, had to be done to ensure that the model was
back in good working condition:
o Non metallic parts of the smoke producing container were burnt and damaged in
the previous tests therefore the whole container was replaced by all metal
(galvanised iron sheet) container. And some ports for tube attachment were
included in the design. Other changes that were made in the course of experiments
are explained in chapter 4.
54
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
o Initially, the smoke extractor used for forcing smoke into the test section was at
the top of the smoke producing container. Due to high temperature the plastic
parts got damaged. To solve this problem, the smoke extractor was relocated
along the smoke delivery tube, as shown in figure 2.3.
o The joints were sealed with silicon sealant to make the test section air tight.
o The blades of the fan for smoke extractor were replaced by a smaller one to
ensure that the smoke enters the test section at a very low velocity. This was an
improvement on the model because smoke entering test section at high velocity
compromises the flow pattern created by the jet fan.
o The tapered end of the through-road was found open, as a result there was almost
no air flowing in the through-road when the fan was switched on. The tapered end
was covered by plastic sheet to ensure air flow.
o Since the air compression system can be used by several persons at the same time,
the flow fluctuated. A pressure regulator20 was fixed to ensure constant supply
pressure. Although, the regulator was not small enough to show the readings
accurately, it was able to regulate a constant pressure. The flow velocity was set
by using anemometer.
20
It was the smallest regulator that was found at that time.
55
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
The steady state simulations were done in order to investigate the air flow patterns in the
underground coalmine.
To facilitate steady state experiment holes with diameter of 6 mm were drilled through
floor of the heading as shown in figure 2.4. Taking the length, height and width of the
scale model as x, y and z direction, the spacing of holes on the model in x-direction was
322 mm and that in z-direction 67 mm, representing 4.83 m and 1 m in full scale scenario
respectively. Since some points above the CM were also examined for air flow, the holes
were drilled on the CM as well. Initially, there were a total of 42 holes made on the floor
starting from column C to I. But after the first type of the experiment was done and
analysed it was observed that more points were required around the heading entrance
hence two more columns (columns A and B) were added.
For the purposes of analysis of experimental data, later on in the chapter, each hole was
given a reference number. For example, the hole at top left corner is A6.
Please take note that there is no hole at location A1. This is so because the pipe carrying
from the compressor to the nozzle passed over this position and it was not possible to
have a smoke out at the location. However, this does not affect the investigation much.
56
CHAPTER 2:PROBLEM STATEMENT AND METHODOLOGY
Three different points of mining progression in the heading were selected. See the plan
views of mine geometries in figure 2.5. A single geometry could be enough but three
were deliberately done in order to increase the confidence in the results of the research.
The first one (A), at a point when the first and second box cuts21 are completed, in other
words half the heading has been mined. Secondly (B), at a point where the CM is doing
the third box cut with coal seam 10 m long (equivalent to 666 mm in the scale model)
remaining. And finally (C), at a point where CM is doing the last box cut with coal seam
7 m long (equivalent to 467 mm in the scaled model) remaining. These three points,
figure 2.5, are referred to as MG-1, MG-2 and MG-3, respectively. MG stands for Mine
Geometry.
21
Box cuts are described under bord and pillar mining techniques.
57
CHAPTER 2: PROBLEM STATEMENT AND METHODOLOGY
(A) MG-1
(B) MG-2
(C) MG-3
Figure 2.5: Mine Geometries Shown in plan view Used in the Steady State Scaled
Experiments (from the top MG-1, MG-2 and MG-3).
58
CHAPTER 2: PROBLEM STATEMENT AND METHODOLOGY
In all the three scale mine geometries the jet fan position is the same, i.e. 80 mm into the
heading, 67 mm from the sidewall and 47 mm above the floor. This scaled position
represents the actual position used in most mines in RSA.
The objective of this experiment was to find out the time taken to extract the air out of the
test section. To achieve this objective smoke was also utilize as in steady state
experiments. Unlike the steady state experiment where smoke was used to assist in
visualizing the air movement, in this instance smoke was used to represent air in the
actual mine that is polluted with dust and mine gas. Smoke was selected to represent
pollutants because its presents could be detected by sensors. Fresh air was then let into
the test section to clear out the polluted air. Time taken for the smoke to completely be
removed at different locations was recorded. The locations of sensors are described in the
next section but details of the sensors and conduction of unsteady state experiment are
described in chapter 4.
Sensors were inserted in the scaled-down mine model at several location of interest as
shown in figure 2.5. CAD geometry representing the set up in figure 2.5 was also
prepared for the purposes of simulation and comparison with experimental results.
The numerical positions for the sensors are given in table 2.1. The origin of these
positions is at right22 bottom corner of the heading entrance of mine geometry as
indicated in figures 2.6. These dimensions are read to the centre of the sensors.
22
Right hand side of a person standing at heading entrance and facing the blind end of heading.
59
CHAPTER 2: PROBLEM STATEMENT AND METHODOLOGY
60
CHAPTER 2: PROBLEM STATEMENT AND METHODOLOGY
Sensors S1, S2, S3 and S4 were located at the blind end since this is the most problematic
area as far as mine ventilation is concerned. Sensors S5, S6 and S7 were located midway
the heading while S8 was located at the heading entrance. This was deliberately done to
ensure that almost all the heading space was covered.
61
CHAPTER 3: CFD MODEL AND CFD RESULTS
3.1 Introduction
As it has already been mentioned in previous chapter, CFD modelling was used in this
investigation. The importance of correct CFD model is immeasurable, as an incorrect
model or non repeatable solution will render the optimization process useless.
In this study the CFD software, Star CCM+ version 3.02.006 was used. Due to the fact
that the software was used for the first time by the researcher and to ensure the
correctness of the model; a few simple internal fluid flow problems were done using both
analytical and computer simulation methods and then the results were compared. This
was done for the purposes of building confidence in the use of the Star CCM+ software
before tackling the actual research problem.
The first section of this chapter gives basic concepts of internal fluid flow. Then the
theoretical and simulated pressure change solutions of fluid flow in selected closed
conduit23 were compared as follows: flows in constant diameter pipe and rectangular
duct, duct with 90 0 mitre bend (without vanes), pipe and duct with long sweep elbow and
pipe with an orifice. The analytical solutions were solved using Microsoft excel. The last
section in this chapter presents CFD models of scaled down model of the underground
coalmine.
In order to understand how the solutions to fluid flows in the closed conduits were
arrived at, some of the basic concepts of fluid flow are discussed first.
23
Both pipes and ducts are referred to as conduits.
62
CHAPTER 3: CFD MODEL AND CFD RESULTS
The flow of a fluid in a conduit may be laminar flow or it may be turbulent flow. In a
laminar flow the fluid flows with no significant mixing of neighbouring fluid particles. If
a dye were injected into the flow, it would not mix with the neighbouring fluid except by
molecular activity; it would retain its identity for a relatively long period of time. Viscous
shear stress always influences a laminar flow.
In a turbulent flow fluid motions vary irregularly so that quantities such as velocity and
pressure show a random variation with time and space coordinates. The physical
quantities are often described by statistical averages. A dye injected into a turbulent flow
would mix immediately by the action of the random moving fluid particles; it would
quickly lose its identity in this diffusion process [21-22].
The flow regime depends on three physical parameters describing the flow conditions.
The first is the length scale of the flow field, such as the diameter of the pipe. If this
length is sufficiently large, a flow disturbance may increase and the flow may be
turbulent. The second parameter is a velocity scale; for a large enough velocity the flow
may be turbulent. The third parameter is the kinematic viscosity; for a low enough
viscosity the flow may be turbulent.
The three parameters can be combined into a single parameter that can serve as a tool to
predict the flow regime. This quantity is called Reynolds number24, a dimensionless
parameter, defined as
24
It is named after Osborne Reynolds (1842 – 1912) who did a lot of work in this area.
63
CHAPTER 3: CFD MODEL AND CFD RESULTS
ρVD
Re = (3.1)
μ
Where D and V are a characteristic length and average velocity, respectively; for example
in a pipe flow D could be diameter of the pipe and V the average velocity of the fluid. ρ is
density and μ is viscosity25 due to linear deformation. Since the viscosity is often divided
by the density in the derivation of equations, it has become useful and customary to
define kinematic viscosity represented by symbol υ.
μ
υ = (3.2)
ρ
VD
Re = (3.3)
ν
The flow can also be intermittently laminar and turbulent. This phenomenon occurs
during the transition from laminar to turbulent when the Reynolds number is close to
critical Reynolds number, Re crit. For conduit flow the value of Reynolds number must be
less than approximately 2100 for laminar flow and greater than approximately 4000 for
turbulent flow [22].
25
Viscosity can be thought of as the internal stickiness of a fluid [21].
26
Steady flow: flow in which physical quantities do not change with time.
27
Unsteady flow: time dependent flow where physical quantities change with time.
64
CHAPTER 3: CFD MODEL AND CFD RESULTS
3.2.1.2. Pressure
Fully developed steady flow in a constant diameter pipe may be driven by gravity and/or
pressure forces. For horizontal pipe flow, gravity has no effect except for a hydrostatic
pressure variation across the pipe that is usually negligible. It is the pressure difference,
Δp = p1 – p2, between one section of the horizontal pipe and another which forces the
fluid through the pipe. Viscous effects provide the restraining force that exactly balances
the pressure force, thereby allowing the fluid to flow through the pipe with no
acceleration.
The need for pressure drop can be explained using two different points of views. In terms
of a force balance, the pressure force is needed to overcome the viscous forces generated.
In terms of an energy balance, the work done by the pressure force is needed to overcome
the viscous dissipation of energy throughout the fluid [21-23].
Dρ
= 0 (3.4)
Dt
Among other fluids, low-speed gas flows, such as the atmospheric flow are also
considered to be incompressible flows. The Mach number28, M is useful in deciding
28 V
The Mach number, named after Ernest Mach (1838 – 1916), is defined as M= where V
c
is the gas speed and the wave speed c = kRT [21].
65
CHAPTER 3: CFD MODEL AND CFD RESULTS
Head loss is probably the most calculated quantity in conduit flow. If the head loss,
expressed in metres, is known in a developed flow, the pressure change can be calculated.
The head loss that results from the wall shear in a developed flow is related to the friction
factor by the Darcy-Weisbach equation, namely,
hL = f
(3.5)
If the friction factor is known, the head loss and the pressure change can be found. In
equation 3.5, it can be observed that the head loss is directly proportional to the average
velocity of the fluid.
29 8τ
Friction factor is a dimensionless wall shear defined by f=
ρV ^ 2
66
CHAPTER 3: CFD MODEL AND CFD RESULTS
The friction factor f depends on the various quantities that affect the flow, written as
f = f (ρ, μ, V, D, e) (3.6)
ρVD e
f = f( , ) (3.7)
μ D
e
Where, is the relative roughness.
D
ρVD
and is Reynolds number.
μ
For, non circular cross sections, good approximations can be made for the head loss by
using the hydraulic radius R, defined by:
A
R = (3.8)
P
Where A is the cross sectional area [m2] and P is the wetted perimeter [m], that perimeter
where the fluid is in contact with the wall of the conduit. Therefore, the Reynolds
number, the relative roughness and the head loss for non circular conduits are as follows:
4 ρVR 4VR
Re = or (3.9)
μ ν
67
CHAPTER 3: CFD MODEL AND CFD RESULTS
e
Relative roughness = (3.10)
4R
L V2
hL = f (3.11)
4R 2g
Experimental data that relate the friction factor to the Reynolds number have been
obtained for fully developed pipe flow over a wide range of wall roughness. The results
of these data are presented in Appendix C, which is commonly referred to as the Moody
diagram30
Apart from head losses due to a developed flow in a conduit there are also additional
losses that are referred to as minor losses, even though such losses can exceed the
frictional losses. These losses are due to fittings and other conduit geometries which
include: valves, elbows, enlargements, contractions, orifice, inlets, outlets, bends etc.
Each of these devices causes a change in the magnitude and/or the direction of the
velocity vectors and hence results in head loss. In general, if the flow is gradually
accelerated by a device, the losses are very small; relatively large losses are associated
with sudden enlargement or contractions because of the separated31 regions that result.
V2
hL = K (3.12)
2g
30
Named after Lewis F. Moody (1880 – 1953) [21]
31
A separated flow occurs when the primary flow separates from the wall.
68
CHAPTER 3: CFD MODEL AND CFD RESULTS
Values of K have been determined experimentally for the various fittings and geometry
changes of interest in piping systems. One exception is the conical enlargement and
contraction from diameter D1 to diameter D2, for which the value of K can be calculated.
Since an orifice is a combination of contraction and enlargement it can also found by
their summation. The formulae for the three exceptions are as follows:
( )
2
2.6 sin θ ⎡1 − D1 ⎤
2
2 ⎢⎣ D 2 ⎥⎦
If θ ≤ 450, Ke= (3.13)
D1 (
D2
4
)
( )
2
⎡1 − D1 2
⎤ sin θ
⎢⎣ D 2 ⎥⎦ 2
If 450 < θ ≤ 1800 Ke= (3.14)
D1(D4
4
)
69
CHAPTER 3: CFD MODEL AND CFD RESULTS
( )
2
⎡1 − D1 2
⎤
⎢⎣ D2 ⎥⎦
And Ke= (3.15)
(
D1
D2
4
)
0
0.8 sin θ ⎡1 − D 2
2 ⎢⎣
2
⎤
D1 ⎥⎦
( )
If θ ≤ 45 , Kc= (3.16)
D2 (D1
4
)
⎢⎣
(
0.5⎡1 − D 2 ⎤ sin θ
D1 ⎥⎦ 2
) 2
(
0.5⎡1 − D 2
⎢⎣
2
⎤
D1 ⎥⎦
)
And Kc= (3.18)
(D2
D1
4
)
70
CHAPTER 3: CFD MODEL AND CFD RESULTS
For an orifice
Orifice is the combination of sudden contraction and enlargement, therefore its loss
coefficient K c, is the sum of the two losses.
Ko = Kc + Ke
A bend has a relatively large loss coefficient. This results primarily from the secondary
flow caused by the fluid flowing from high-pressure region to the low-pressure region. In
addition, a separate region occurs at the sharp corner. Energy is required to maintain a
secondary flow and the flow in the separate region. This wasted energy is measured in
terms of the loss coefficient.
The loss coefficients for various geometries are presented in Table 3.1
71
CHAPTER 3: CFD MODEL AND CFD RESULTS
Diameter 2.5 cm 5 in 10 cm 5 cm 10 cm 20 cm
Globe valve (fully open) 8.2 6.9 5.7 8.5 6.0 5.8
(half open) 20 17 14 21 15 14
(one-quarter open) 57 48 40 60 42 41
Angle valve (fully open) 4.7 2.0 1.0 2.4 2.0 2.0
Swing check valve (fully open) 2.9 2.1 2.0 2.0 2.0 2.0
Gate valve (fully open) 0.24 0.16 0.11 0.35 0.16 0.07
Return bend 1.5 0.95 0.64 0.35 0.30 0.25
Tee (branch) 1.8 1.4 1.1 0.80 0.64 0.58
Tee (line) 0.9 0.9 0.64 0.39 0.30 0.26
Standard elbow 1.5 0.95 0.23 0.30 0.19 0.15
Long sweep elbow 0.72 0.41 0.23 0.30 0.19 0.15
450 elbow 0.32 0.30 0.29
Square-edged entrance 0.5
Re-entrant entrance 0.8
Well-rounded entrance 0.03
Pipe exit 1.0
Sudden contraction33 Area ratio
2:1 0.25
5:1 0.41
10:1 0.46
Orifice Area ratio A/A o
1.5:1 0.85
2:1 3.4
4:1 29.0
≥ 6:1 2.78(A/Ao-0.6)2
32
Values for other geometries can be found in Technical Paper 410, Crane Company, 1957.
33
This is based on exit velocity V2.
34
This is based on entrance velocity V1.
72
CHAPTER 3: CFD MODEL AND CFD RESULTS
Before presenting the pressure changes for selected geometries description of steps that
were followed when calculating and simulating pressure change in a conduit are
presented.
o Calculate the Reynolds number. Re is found using equation (3.1) if the conduit is
a pipe or equation (3.8) if a conduit is a duct. To use any of the two equations one
needs to know the cross section dimensions of the conduit, the velocity of the
fluid flowing in the conduit. The density, viscosity or kinematic viscosity can be
extracted from the table of fluid properties if the temperature of the fluid is
known. The properties of water and air required in this research are attached in
Appendix D.
o Afer Re of the flow and relative roughness of the conduit surface are calculated.
o Calculate the fluid head resistance to overcome the flow through the conduit.
Equations 3.5 or 3.11 are used if the conduit is a pipe or a duct respectively.
73
CHAPTER 3: CFD MODEL AND CFD RESULTS
o Determine loss coefficient K, for the fitting(s) within the conduit. The nominal
loss coefficients for common commercial fittings are shown in table 3.1 and
formula for pipe enlargement; contraction and orifice are given in equations 3.13
to 3.19.
o Calculate the total head resistance or loss that is present in the conduit.
o Calculate the pressure loss in the conduit. This is found by applying the following
formulae:
ΔP = ρgh = γh (3.20)
Before one starts simulation process in Star CCM+, the geometry in/on which the flow
will be simulated must be done first. This can be achieved using drawing and design
software like Solid Works, Auto CAD, etc. The geometries are then saved in file types
that star CCM+ can import from. These files are [20]:
In this investigation geometries were created in Solid Works then saved in a Stereo
lithography file.
After the necessary geometry is created the simulation process in Star CCM+ proceeds as
follows [20]:
This section presents the comparative results for pressure changes in different conduits
found through analytical and simulation methods. All the conduits under consideration
35
A region is a volume domain (or surface in a two-dimensional case) in space and has to be completely
surrounded by boundaries.
36
Meshed geometries prepared in pro-star, gridgen, fluent, gambit, star-CD, ICEM are compatible with
Star CCM+ [20].
75
CHAPTER 3: CFD MODEL AND CFD RESULTS
are assumed to be made of smooth surfaces37 and are lying in a horizontal position. Both
the analytical and numerical pressure changes are plotted on the same graph for ease of
comparison. An analytical calculation for a pipe with an orifice only is shown in
Appendix E as an example.
Water flow in a constant diameter straight pipe with the following details was simulated
and then pressure drop was analytically done:
The pressure change for both simulation and analytical calculation are shown in the graph
in figure 3.3.
37
The friction factors are taken from the line indicated as smooth pipe in Moody Diagram on appendix C.
76
CHAPTER 3: CFD MODEL AND CFD RESULTS
1400.0
1200.0
1000.0
Pressure (Pa)
800.0
600.0
400.0
200.0
0.0
0 0.1 0.2 0.3 0.4 0.5
Distance (m)
Theoretical Simulation
The two plots for theoretical and simulated pressure drop are almost the same. The small
difference, approximately 50 Pa which, translates to 4% difference, may be due
inaccurate reading from Moody diagram.
Water flow in a square straight duct with the following details was simulated and then
pressure change was analytically done:
Length 1.00 m
Inlet velocity 1.00 m/s
The pressure change for both simulation and analytical calculation are shown in the graph
in figure 3.4.
300
250
200
Pressure (Pa)
150
100
50
0
0 0.2 0.4 0.6 0.8 1
Distance (m)
Theoratical Simulated
Figure 3.4: Pressure Change in Square Duct by Simulation and Analytical Calculation
Just like in the constant diameter pipe simulation two plots for theoretical and simulated
pressure drop in the straight duct are also almost the same with a difference of about
10%. The difference may arise from inaccurate reading from Moody diagram.
Water flow in a rectangular duct with 90° mitre bend, figure 3.5, having the following
details was simulated and then pressure change was analytically done:
78
CHAPTER 3: CFD MODEL AND CFD RESULTS
The pressure change for both simulation and analytical calculation are shown in the graph
in figure 3.6.
79
CHAPTER 3: CFD MODEL AND CFD RESULTS
900.0
800.0
700.0
Pressure Drop
600.0
500.0
400.0
300.0
200.0
100.0
0.0
0 0.5 1 1.5 2
Axial Length (m)
Theoretical Simulation
Figure 3.6: Pressure Change in a rectangular duct with 90 0 mitre bend by Simulation
and Analytical Calculation
The pressure drops from the two results have a difference of approximately 150 Pa but
the profiles of the two plots are similar. The main reason for the difference in plot
profiles could be the position where the minor head loss occurs. Theory assumes that it
occurs right at the bend while in real practice it occurs after the bend.
Water flow in a pipe with long sweep elbow, figure 3.7, having the following details was
simulated and then pressure change was analytically done:
80
CHAPTER 3: CFD MODEL AND CFD RESULTS
Length 1.50 m
Inlet velocity 2.00 m/s
Figure 3.7: Long sweep elbow with straight pipe on both ends
The pressure change for both simulation and analytical calculation are shown in the graph
in figure 3.8.
81
CHAPTER 3: CFD MODEL AND CFD RESULTS
700
600
Pressure Drop (Pa)
500
400
300
200
100
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Distance (m)
Theoretical Simulation
Figure 3.8: Pressure Change in a pipe with long sweep elbow by Simulation and
Analytical Calculation
The pressure change found by the two methods is almost the same. The only difference is
the gradient which might have been caused by point where the minor loss is assumed to
occur in the theoretical calculation.
3.2.3.5. Orifice
Water flow in a pipe with an orifice, figure 3.9, having the following details was
simulated and then pressure change was analytically done:
82
CHAPTER 3: CFD MODEL AND CFD RESULTS
The pressure change for both simulation and analytical calculation are shown in the graph
in figure 3.9.
83
CHAPTER 3: CFD MODEL AND CFD RESULTS
1400
1200
Pressure Drop
1000
800
600
400
200
0
0 0.5 1 1.5 2
Axial Length (m)
Theoretical Simulation
Figure 3.10: Pressure Change in a pipe with orifice by Simulation and Analytical
Calculation
It can be observed that pressure change is similar with a slight difference in the point
where the drop due to orifice starts to occur. This is because the theory assumes that the
drop starts exactly at the orifice itself. The difference in pressure drop is only 3% of
theoretical one.
In order to achieve the objective of this investigation, two types of numerical solutions
were done using a scaled down underground coalmine. The first one was a steady state
condition and the other unsteady state condition. This section discusses the two types of
numerical solutions, starting from development of geometries in CAD, through to the
84
CHAPTER 3: CFD MODEL AND CFD RESULTS
results. As explained earlier, the results were later on validated with experiments whose
details are covered in chapter 4.
This section describes in details steps that were taken in order to arrive at the steady state
numerical solutions.
The geometry and other parameters that were used for simulation were 15% smaller than
the size and parameters prevailing in the actual mine. This was done because the
experiments to validate the simulations were done on a 15%-scale model of the
underground coalmine. The scaling down of mine dimensions and parameters like air
velocities are explained in chapter 2.
Solid Works CAD software was used to develop the geometry. The geometry comprising
the heading, through-road, jet fan, scrubber and CM was developed in one block. The
shape of the CM was just cut in the block in order to minimize the size of the actual
volume to be meshed which translates into less number of cells therefore reduced
meshing and iteration time. The fan was inserted into the block through assembling. See
figure 3.11. The geometry was then stored in Stereo lithography format ready to be
imported by star CCM+.
85
CHAPTER 3: CFD MODEL AND CFD RESULTS
The geometry was imported to star CCM+ from solid works as one region and one
boundary. Therefore it was necessary to split the geometry so that Star CCM+ should be
able to recognize the fan, the tunnel, CM and scrubber. The boundaries were split,
renamed and relevant boundary types were specified accordingly. The units of
importation were millimetres the same as the one used when the geometry was developed
in solid works. The geometry was imported with surface mesh too coarse to give accurate
simulation results. Therefore the next step was to re-mesh the surface geometry before
volume meshing was done.
As it has already been explained earlier, the volume mesh in a simulation is the
mathematical description of the space (or geometry) of the problem being solved. The
86
CHAPTER 3: CFD MODEL AND CFD RESULTS
+
volume mesh in Star CCM can contain either trimmed, polyhedral or tetrahedral type
cells for each mesh region, determined via the selection of mesh models. Additionally,
prism layers38 can be automatically included next to wall boundaries and interfaces.
A grid is generated in a domain where the fluid flow equations are solved through
computational techniques. To obtain good results the appropriate grid must be applied.
This section briefly, gives a brief description of the three different types of volume mesh
models, explains mesh or grid independence, describes how a decision to use a specific
model was made and gives the reference values that were used in the mesh model.
Tetrahedral meshes provide an efficient and simple solution for complex mesh generation
problems. Out of the three, it is the fastest and uses the least amount of memory for a
given number of cells.
Polyhedral meshes provide a balanced solution for complex mesh generation problems.
They are relatively easy and efficient to build, requiring no more surface preparation than
the equivalent tetrahedral mesh. They also contain approximately five times fewer cells
than a tetrahedral mesh for a given starting surface.
38
A prism layer mesh is composed of orthogonal prismatic cells that usually reside next to wall boundaries
in the volume mesh. They are required to accurately simulate the turbulence and heat transfer.
87
CHAPTER 3: CFD MODEL AND CFD RESULTS
The trimmer mesh provides a robust and efficient method of producing a high quality
mesh for both simple and complex mesh generation problems. The model is not directly
dependent on the surface quality of the starting surface and as such is more likely to
produce a good quality mesh for most situations.
In terms of general accuracy for a given number of cells, the trimmer and polyhedral cell
type meshes always produce the most accurate solution when compared to tetrahedral
mesh [Star CCM+ Notes].
When solving CFD problems it is important to ensure that the solutions are not affected
by the size of the mesh or grid. The ability to retain consistency across varying cell sizes
is called mesh independence. Mesh independence can be achieved by taking three runs of
the same problem but on very different mesh size. If the solutions from the three runs
show little variation then the mesh is deemed to be independent.
In this investigation cell size of 0.1 m was utilized because mesh independence was
observed around this size. The details on how the mesh was sized are given in the next
section.
Before Star CCM+ was commanded to mesh the geometry reference values had to be
specified. Reference values determine the size, number and quality of cells among other
things. Reference values that were used to mesh the models are shown in table 3.2.
88
CHAPTER 3: CFD MODEL AND CFD RESULTS
Table 3.2: Reference values that were employed to mesh the model
Parameter Value
Base value 0.1m
Number of prism layers 1.0
Prism layer thickness (Relative size) 5.0
Surface curvature 36
Surface growth rate 1.3
Surface proximity (# points in gap) 2.0
Surface size
o Relative minimum size 0.1
o Relative target size 100
Tet/poly density
o Density 1.0
o Growth Factor 1.0
Tet/poly source blending 1.0
The import mesh had 520 faces and the numbers of cells for each mesh model are given
in table 3.3.
Table 3.3: Number of Cells for MG 3 for the Three Mesh Models
89
CHAPTER 3: CFD MODEL AND CFD RESULTS
The model that was used to define the physical phenomenon of steady state solutions is
as shown in table 3.4:
90
CHAPTER 3: CFD MODEL AND CFD RESULTS
Monitors enable summary information from the simulation to be sampled and saved
during the solution. This information can then be displayed in an XY plot while the
solution is progressing or post-processed after iteration or unsteady time-stepping is
complete. They are useful for watching the behaviour of numerical or engineering
quantities as the solution evolves, and can be used to define stopping criteria that can stop
the solution iteration.
A report presents a computed summary of the current simulation or CPU data. The report
summaries are useful for post-processing, and enable engineering quantities to be
computed.
After setting monitors and reports then the stopping criteria and solvers are set before
initiating iteration process.
Unsteady state simulations were done in order to investigate the rate of extracting an
initial volume of air in the underground coalmine model. This was later used to determine
the optimum position of jet fan for different mine heading lengths in chapter 6.
39
The velocity value of 10 m/s was used instead of 9.42 m/s because it was difficult to accurately set this
velocity.
91
CHAPTER 3: CFD MODEL AND CFD RESULTS
In the simulation, it was assumed that the test section was initially filled with smoke
simply because smoke was used in the verification experiment in chapter 4. Then the jet
fan was switched on to bring in flesh air, thereby removing the smoke. The smoke
represents the volume of air in the actual mine that is polluted with dust and mine gas as
it has already explained in the methodology in chapter 2.
The Multi-Component Gas model was used for this analysis and results were presented in
form of XY- plot of Mass Fraction of Air against Time.
Most of the steps that were done to get unsteady state solution were the same as those for
steady state. Only those steps that were different are explained in this section.
Solid Works CAD software was used to develop the geometry. Geometry MG 3, figure
2.5, was utilized for unsteady state simulation.
For unsteady state a non-reacting multi-component gas and implicit unsteady were
selected for material and time models respectively. The initial condition of mass fraction
was set as (0.0, 1.0).40
The model that was used to define the physical phenomenon of unsteady state solutions is
shown in table 3.6.
40
This implies that there was no fresh air in test section but it was completely filled with smoke.
92
CHAPTER 3: CFD MODEL AND CFD RESULTS
41
This implies that the jet fan is blowing 100 percent fresh air. This is also the same with the air flowing
into the through-road.
93
CHAPTER 3: CFD MODEL AND CFD RESULTS
An implicit unsteady solver of 1.6 seconds and a maximum inner iteration of 1 were
found to produce results similar to the experimental results.
Mass fraction of air was monitored on averaged area of sensor boundaries and the results
were plotted against time.
3.5. Results
Results for MG 3 are presented in details in this section and those for MG 1 and MG 2
are in Appendix F. Star CCM+ is capable of presenting results in many different forms
including vectors. Streamline diagrams are given here as they give a better representation
of the flow field than would have been the case with vector plots. Figure 3.12 shows the
results in orthographic and isometric projections.
94
CHAPTER 3: CFD MODEL AND CFD RESULTS
Front Elevation
End Elevation
Plan
Isometric View
Figure 3.12: Flow Pattern of Air in Mine Geometry 3 Presented in Orthographic and
Isometric Views
95
CHAPTER 3: CFD MODEL AND CFD RESULTS
The CFD model results for the rate of extracting smoke out of the test section for the
mine geometries MG-3, using tetrahedral and polyhedral volume mesh models are shown
in graph format in figures 3.13 and 3.14. The reason for using tetrahedral and polyhedral
volume meshes was to assist in choosing the best volume mesh to be utilized for the rest
of numerical analysis. A decision as to which type of mesh to be used for the rest of
numerical analysis was made after comparing the results for both types with experimental
results.
1.0
0.8 Sensor1
Mass Fraction of Air
Sensor2
0.6 Sensor3
Sensor4
0.4 Sensor5
Sensor6
0.2 Sensor7
Sensor8
0.0
0 50 100 150 200 250 300
Time {s}
Figure 3.13: Graph of Mass Fraction of Air against Time for MG-3, Tetrahedral Mesh
Model.
96
CHAPTER 3: CFD MODEL AND CFD RESULTS
1.0
0.8
Sensor1
Mass Fraction of Air
Sensor2
0.6
Sensor3
Sensor4
0.4 Sensor5
Sensor6
0.2 Sensor7
Sensor8
0.0
0 100 200 300
Time {s}
Figure 3.14: Graph of Mass Fraction of Air against Time for MG-3, Polyhedral Mesh
Model.
For both mesh models, it can be observed that a graph for sensor 8 takes off first,
followed by sensors 5, 6 and 7 and lastly, sensors 1, 2, 3 and 4. This means that
extraction of initial volume of air began at sensor location 8, followed by locations 5, 6
and 7 and lastly, 1, 2, 3 and 4. This was so because location 8 was closer to the fan than
the rest while sensor 5, 6 and 7 were closer than 1, 2, 3 and 4. Therefore, fresh air from
the fan reached sensor 8 first before the rest, etc.
Graphs for sensors 5, 6, 7 and 8 rose sharply soon after taking of then after approximately
50 s their gradients reduces drastically. The slower rate of extraction beyond the 50 s
mark could be due to the effects of recirculation and diffusion rate. For sensors 1 to 4 in a
polyhedral mesh model the rise before the 50 s mark is not as sharp as the rest. This could
be because they are located at the blind end of the heading where the air jet could not
penetrate with much strength.
97
CHAPTER 3: CFD MODEL AND CFD RESULTS
The major difference between tetrahedral and polyhedral mesh models is in the way the
graphs for sensors located at blind end of the heading (i.e., sensors 1, 2, 3 and 4) rose.
The gradient for polyhedral mesh model was slower than that for tetrahedral. The
difference comes about because of the difference in the design of the two volume meshes.
Polyhedral mesh model was chosen because it discriminated the behaviour of the graphs
of sensors far much better than tetrahedral mesh model, especially between the one
located at blind end and the rest. Secondly, when the graphs for the two mesh models
were compared to the similar graphs for experimental results, the trend of the graphs for
polyhedral mesh model were found to be closer than those for tetrahedral mesh model.
98
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
4.1 Introduction
The correctness of CFD model depends upon correct meshing and physical inputs.
Otherwise, one might end up with a simulation that runs and converge very well but gives
wrong results. For this reason, it was decided that experimental verification should be
conducted in order to ensure the correctness of the CFD model that were constructed in
chapter 3, before further numerical solutions were done.
Two types of experiments were conducted. The first was the steady state (time
independent) and the other unsteady state (time dependent).
The methodology that was used to carry out investigation for this study was covered in
chapter 2. This chapter describes how experiments were conducted on the 15%-scale
model and reveals the results. The experimental results are compared with numerical
solutions later in chapter 5.
The first type of experiment was the steady state. The objective of this experiment was to
find out the pattern of air flow in the test section. As it has already been mentioned in
chapter 2, smoke was utilized to visualize air flow. The flow patterns at each smoke
outlet hole were recorded by a visual camera. This was done in order to facilitate repeated
observation of flow direction.
99
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
In steady state experiments, the flow patterns of the air in the test section were recorded
so that it could be analysed in detail at a later stage. A digital camera with a resolution of
was used to record video clips. For each position in the test section two video clips were
taken, one from the transparent heading side of the model and the other from the top.
Positions for recording the video clips were fixed to ensure that images captured could be
accurately checked for repeatability. To achieve this the camera was mounted on a tripod
stand and positions for tripod stand were carefully setup according to predefined
markings on the floor of the lab as shown in figure 4.1. These are the positions that were
used to capture images from the side. Similarly, positions on the camera mounting rail
above heading of the model were also marked.
Figure 4.1: Marked positions for tripod stand for the camera
100
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
When images were being captured all the lights in the laboratory were switched off and a
flow observation light on the scale model was switched on in order to maximise
visualisation of smoke.
The following is the procedure that was utilized in process of conducting the experiment:
o The valve on the compressed air system was opened to give air velocity at the
nozzle exit of 10 m/s.
o The smoke tubes (labelled 1 in figure 2.3) were inserted into three consecutive
holes along the width of the heading and set at same level or height at a single
time. The heights that were used are 47 mm, 165 mm and 250 mm, measured
from floor level of the heading. These heights are referred to as level 1, 2 and 3
respectively, in the dissertation. This was done to ensure that flow directions at as
many positions as possible in the test section are captured. The rest of the holes on
the floor were sealed off by insulation tape to avoid air leakages.
o A digital camera was set to capture data in form of video clips. Image capturing
process is explained in details in section 4.2.1.1.
o The smoke system, the through-road fan, observation light and scrubber were all
switched on while laboratory lights were switched off.
o As soon as enough smoke started coming out of the tubes in the test section the
camera was switched on to start capturing images. Two video clips were taken for
each position in the test section, one from the side and the other from the top.
Most of video images looked similar therefore a log was kept in order to avoid
confusing them during analysis.
101
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
o After capturing data at one position the three tubes were set again at a different
level or next set of holes. And, the third to fifth steps were followed again until all
the designated points in the test section were done. It should be noted that some
points especially those at level 1 were obstructed by the CM. At those particular
holes only points at levels 2 and 3 were captured.
The second type of experiment was the unsteady state also referred to as transient state.
The objective of this experiment was to find out the time taken to extract the air out of the
test section. In this section, air presence measuring system, experiment set up and
procedure followed in carrying out the experiment are described.
To determine the time taken by air to be extracted out of the test section a measuring
system was devised.
In order to understand the experiment set up better the general framework of the
measuring system is described first. The general framework consists of three stages as
shown in figure 4.2 and are described below [27].
102
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Sensors and transducers convert a physical quantity into an electrical quantity. Therefore,
primary function of the first stage is to detect or to sense the measurand42.
The purpose of the second stage is to modify the transduced information so that it is
acceptable to the third, or terminating stage. The most common function of this stage is to
increase either amplitude or power of the signal, or both, to the level required to drive the
final terminating device. In addition, it may perform other operations, such as selective
filtering to remove noise, integration, differentiation, or telemetering, as may be required.
42
Measurand is the physical parameter being observed and quantified or an input quantity to the measuring
process [27].
103
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The third stage provides the information sought in a form comprehensible to humans or
to a controller. Graphs of resistance against time are utilized for this purpose in this
experiment. These graphs were produced on the computer using Microsoft Excel.
Figure 4.3, shows the block diagram for the measuring system that was designed for this
experiment and the details of each stage are described below.
Figure 4.3: Block diagram of the specific measuring system for unsteady state
experiment
4.2.2.1.1. Sensor-transducer
Light-detecting transducer called Retro sensor was used to detect the measurand, which
in this case was smoke. The retro sensor detects the presence of an object by generating
light and then looking for its reflectance from the object to be sensed. Since air which is
the sensed object does not reflect light smoke had to be injected into it to serve this
purpose. A retro sensor comprising a phototransistor43, Light Emitting Diode (LED) lamp
and a cardboard structure was made and is shown in figure 4.4.
43
Phototransistor is basically a photodiode followed by one or two stages of amplification incorporated in
the same package to enhance the output. Photodiode are semiconductors that produce current flow when
they absorb light.
104
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The sensor was made as small as possible to minimize the interruption of air flow in the
test section. The surface area for the roof supports was also as small as possible to
minimize the flow interruption above the LED lamp and phototransistor. The under
surface of the roof was painted black (using non gloss paint) to reduce its reflectance
capability.
105
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Agilent 34970A Data Acquisition Unit (DAU) was used in the experiment for signal
conditioning. The signal sent by the retro sensor in form of electrical current (I) was
converted to electrical resistance (R).
Data that was captured was stored in the DAU memory to be uploaded into the computer
for analysis.
A computer was used to readout and analysis the data from DAU. The computer needs to
have relevant software to be able to upload and analyse data from DAU. In this case,
Agilent BenchLink Data Logger44 and Microsoft excel were used for uploading and
analysis respectively.
Unlike in the steady state experiment where several mine volume were used, in the
unsteady state experiment only one volume was utilized.
The exact locations for the sensors have already been explained in chapter 2 and
illustrated in figure 2.5.
44
Agilent BenchLink Data Logger is a windows-based application designed to make it easy to use Agilent
34970A with a computer for gathering and analyzing measurements.
106
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
4.2.2.2.2. Smoker
The smoke producing device that was used in the steady state experiment was found to be
lacking for this experiment. The reason being that the smoke flow rate into test section
was very low and since the test section was not fully airtight it was very difficult and time
consuming to fill the volume with enough smoke to get reasonable data.
Another type of smoke producing device was constructed and represented in figure 4.5. It
was able to reduce the time needed to fill the volume dramatically.
This smoke producing device comprises a coil heating element fixed in a metal tray. The
smoke source which in this case was two pieces of commercial hardboard placed
underneath and on top of the heating element. Then both the element and the hardboard
were buried in sand to avoid a flame. The device was placed almost at the middle of the
test section. A cable was attached to the tray for pulling the device out of the test section.
107
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Before using the sensors for the experiment it was necessary to examine if they were able
to give repeatable results. This was done by simply switching the light in the laboratory
on and off while the DAU was recording the results in ohms. The results for this exercise
shown in the graph, figure 4.6 reveals that the sensors gave repeatable results and could
be used with confidence. The peaks are for light on and valleys for light off.
320
Sensor1
300
Sensor2
280
Resistance {kohm}
Sensor3
260 Sensor4
240 Sensor5
220 Sensor6
200 Sensor7
180 Sensor8
0 20 40 60 80 100
Number of Scans
Figure 4.6: A Graph of Resistance against Time with Laboratory Lights on and off
108
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The sensors had to be calibrated before using them for the experiment so that the
necessary adjustments could be made later on the experiment results. The calibration was
done as follows:
A regulator transformer was used to vary light intensity of LED lamps of Retro sensors.
By adjusting the voltage reading of the regulator transformer the current in the circuit
changed consequently, varying light intensity of LED lamps. Since the transformer was
an Alternate Current (AC) transformer and the LED lamps were Direct Current (DC)
amps, a 3 V adaptor had to be connected to the output of the transformer as shown in the
schematic diagram in figure 4.7.
Figure 4.7: Schematic Diagram of Regulator Transformer, Adaptor and LED Lamps
Circuit.
The regulator transformer was first set at 220 V and the ohms reading for all the sensors
were recorded by DAU. Then the regulator was adjusted upwards at intervals of 5 V up
to 250 V with sensor readings scanned at each interval. To ensure accuracy of calibration
three average readings were recorded at each regulator transformer interval and the
average of the three averages were used in the analysis.
In the analysis the first step was to find the average of all sensors at each particular
voltage value. See table 4.1. The averages were then subtracted from each sensor
109
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
readings to get deviations. Then graphs of deviation against sensor readings were done
for all the sensors. Finally, best line fit of all the graphs were plotted and their formulas
were used for adjusting the experiment data. The process of adjusting experiment data for
sensor 2 is demonstrated in the next section.
Table 4.1: Average of measurements of all sensors taken at different input voltages
Voltage Sensor1 Sensor2 Sensor3 Sensor4 Sensor5 Sensor6 Sensor7 Sensor8 Average
V kohm kohm kohm kohm kohm kohm kohm kohm kohm
220 126.130 317.901 341.616 164.602 110.724 383.385 134.168 130.463 210.249
225 118.389 298.391 317.005 152.744 102.747 348.727 122.039 122.456 197.812
230 112.194 282.777 298.576 143.864 96.7741 323.559 113.232 116.048 185.878
235 105.880 266.862 280.026 134.926 90.7616 298.937 104.615 109.517 173.940
240 102.254 257.725 269.544 129.875 87.3641 285.388 99.8734 105.767 167.224
245 98.416 248.05 258.412 124.512 83.7562 271.29 94.9398 101.796 160.146
250 93.951 236.798 245.653 118.364 79.6205 254.932 89.215 97.1785 151.964
Sensor 2 has been used to demonstrate the adjustment of measured values. Initially, the
average readings of all sensors at each voltage were found as explained in previous
section. Then deviations from the average for the measured value at a particular voltage
were found. Table 4.2 shows how this was done for sensor 2.
110
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Table 4.2: Ohms readings for sensor 2 at different voltage settings, average readings
for all sensors and deviations from the average.
The next step was to plot deviations against ohms readings at each voltage setting. This
was done in order to find the best line of fit. Figure 4.8 shows the said graph. In this case
a linear graph was found to be the best line of fit.
Figure 4.8: Graph of deviation from average reading against sensor readings for
sensor 2 showing best line of fit, its R-squared value and equation.
111
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The next step was now to find the value by which the measured values should be
adjusted. This was achieved by substituting the measured value for x in the equation of
line of fit. Finally, the adjustment value was subtracted from the measured value to get
the final value. Table 4.2 shows the adjustments. The final values for the sensors were
then plotted against time and this is described in experimental results section.
112
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
113
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
114
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The line fit equations for measured values of all sensors are shown in table 4.4 while
ohms readings, deviations from average and the best lines of fit for the other sensors are
shown in Appendix F.
115
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The following activities had to be carried out to complete the unsteady state experiment:
o Connect phototransistor cables to DAU card. Make sure that all the necessary
parameters relevant to this experiment are set. These include: activating the
channels and set them to ohms in order to read resistance, auto range and resetting
the time. Scanning limit was set to continuous and scan interval to 1 second.
o Seal off both the inlet and outlet of the through-road and replace the top cover of
the heading of mine model.
o Switch off laboratory lights to avoid their effect on the experiment and switch on
the LED lamps.
o Press the scan button on DAU to start recording data. Scanning was left to run for
about five minutes before smoke was injected into the test section. This was done
to ensure an initial horizontal graph was achieved. The starting time for scanning
process was recorded for crosscheck purposes.
o After approximately five minutes of scanning the smoker heater was switched on.
The heater was switched off after enough smoke filled the test section. This was
based on the personal judgement. It would have been better to completely saturate
the test section with smoke but the smoke leakage made the environment
uncomfortable so that the experiment had to be done in the shortest time possible.
It was observed that when the heater was switched on, the readings for all sensors
increased significantly by approximately 10 000 ohms and dropped by the same
116
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
margin when it was switched off. Switching on and off time of the heater were
also recorded.
o The smoker was moved to the through-road inlet by pulling the soft wire attached
to smoker tray.
o The jet fan, scrubber fan and the through-road fan were switched on. At the same
time the seals on the through-road inlet and outlet were removed. The smoker was
also taken out of the test section. This marked the beginning of smoke extraction
process. Time at which smoke extraction began was recorded.
o Smoke extraction process was left to run until when all the smoke was removed
from test section. This was determined by the readings on the DAU being
consistent (not fluctuating much), then scanning was stopped.
4.3. RESULTS
The information that was collected in form of video clips was converted into vector
format for them to be easily analysed. Either an arrow or a circle was drawn at each point
where the smoke entered the test section. An arrow indicates smoke direction and a circle
indicate that no definite direction was observed. Please note that the arrows indicate
direction only and not velocity magnitude.
To illustrate how the vector diagrams were constructed from video clips, points E1, E2
and E3 on MG-2 as shown in figure 4.9 and 4.10, representing side view and top view,
respectively.
117
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The air flow represented by the three white streams that are coming out from the tubes
can be seen clearly on the two figures. The arrows are drawn on each stream to indicate
smoke direction.
118
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Figure 4.11 shows the complete vector diagram of MG-3 from the top at levels 1, 2 and 3
and figure 4.12 shows the same vector diagram from the side at planes 1 to 6. The dots
are the positions where the smoke enters the test section from the tubes, the arrows
indicate flow direction that were definite and circles around the dots indicate that the flow
direction could not be determined at all and a circle and an arrow indicate that despite the
smoke circulating it was flowing to a particular direction most of the time. Failure to
determine flow direction at particular point was a result of continuous change of smoke
direction during the whole period of observation in other words there was no dominant
direction of flow. The vector diagrams for MG 1 and MG 2 are in Appendix G.
119
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
Figure 4.11: Vector diagrams developed from video images for mine geometry, MG-3,
at levels 1, 2 and 3, as seen from the top.
120
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
MG 3 Section 1
MG 3 Section 2
MG 3 Section 3
MG 3 Section 4
MG 3 Section 5
MG 3 Section 6
Figure 4.12: Vector diagrams developed from video images for mine geometry, MG-3,
at sections 1, 2, 3, 4, 5 and 6, as seen from the side.
121
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
The raw data whose plots are shown in Appendix H. went through the following process
in order to come up with the final graph shown in figure 4.13.
o Firstly, the effects of electric current fluctuation (see appendix J) when the smoke
heating element was switched on and off were removed. Then all the other
suspected effects of electric current fluctuation due to switching on or off other
electrical machines or devices in the laboratory were also removed.
o Secondly, the data was then adjusted by calibrating measured values, as explained
in 4.2.2.4.
Figure 4.13: Graph showing the concentration of smoke against time around the
sensors in the scaled model.
122
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
After calibrating it was expected that the plots for all the sensors will be close together as
those of sensors 2, 3, 4 and 6. The results in figure 4.14 reveals that something went
wrong with sensors 1, 5, 7 and 8 during the experiment. The probable reasons for this
error could be that conditions under which the calibrations were done might have
changed during of the test. For example, the LED or phototransistor might have shifted.
Since the anomaly was discovered too late while the author way from experiment facility,
it was decided to ignore the malfunctioning sensors in this analysis. Therefore the plot for
normal sensors was done as shown in figure 4.14.
Figure 4.14: Graph showing the concentration of smoke against time around the
sensors in the scaled model after removing malfunctioning sensors.
The removal of these sensors would not affect the investigation so much since validation
process could be done using the remaining four sensors. And since the area of interest
123
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
was the time taken to extract initial air out of test section the measured values of the so
called malfunction sensors produced sensible results. This is covered in chapter 5.
0 – 200 The lines of the plots for all the sensors are almost horizontal. This
was the time between the scan was initiated and before the test
section was filled with smoke.
200 - 400 The lines steeping downwards rapidly. This was the time when the
test section was filled with smoke.
400 - 740 At approximately 400 seconds jet fan and through-road fan were
switched on to start extracting the smoke out of test section. As
soon as the fans are switched on the plots steeps upwards rapidly up
to about 600 seconds mark where the gradient reduces significantly.
740 The plots become horizontal again. This signifies that the smoke is
completely extracted.
From this analysis it can be deduced that it took approximately 360 seconds to
completely remove the smoke out of the test section.
124
CHAPTER 4:EXPERIMENTAL MODEL AND RESULTS
In an ideal case the plots for all the sensors should return back to the starting reading
when all the smoke is extracted out of the test section. This is not the situation with the
experimental results. After the experiment was conducted it was observed that a brownish
thin substance45 had coated the LED lamps, phototransistors and every surface that came
into contact with the smoke. The thin coating reduced the light intensity from LED lamp
and that penetrating the phototransistor. This was the reason the graph could not return to
the starting value.
45
This substance is believed to be a by-product of hardboard combustion.
125
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
5.1 Introduction
The main objective of this work was to demonstrate the possibility of solving coal dust
and mine gas concentration problems in underground coalmine by using CFD modeling.
The first step towards this goal was to model the standard46 method of ventilation in the
coalmines. The CFD model was then validated with experimental results done on scaled
down coalmine model.
The validation was done by comparing the numerical and experimental solutions.
Similarity between the two analysis means that the CFD results can be trusted. And
further numerical analysis can be carried out with confidence.
In order to gain more confidence in numerical analysis both the steady state and unsteady
state results of the two models were compared. Steady state results were done first,
followed by the unsteady state.
Two unsteady solutions were done using tetrahedral and polyhedral types of volume
mesh the comparison of these results against experimental results assisted in choosing the
mesh to be used for the rest of numerical analysis. The polyhedral model produced results
that were closer to the experiment than the tetrahedral meshed model hence the
polyhedral was utilized for the rest of numerical analysis.
46
In the standard method the jet fan is positioned 1.2 m from the entrance, 1 m from the side wall and 0.7
m high.
126
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
Although the experimental results in chapter 4 are presented in plan and front views only,
the end view is interpolated from the two so that a three dimension comparison analysis
is possible. The steady state results in chapter 4 were summed up first before the
comparison process could be carried out. The results for mine geometry 3 (MG 3) are
presented in detail in this section while those for MG1 and MG2 are given in Appendix J.
Please take note that vectors or streamline for scrubber are not included in steady state
analysis. This is because the experimental views do not include them since it was difficult
to visualize them.
The first step in this analysis was to sum up the three top views and the six side views as
seen in vector diagrams, figures 4.13 and 4.14 to a single top view and single side view,
respectively. Then the interpolation of the second side view was done before the
comparison process could be done. To achieve this it was imaged that the three levels of
the top views vector diagrams were transparent and placed directly on top of each other
to come up with one top view vector diagram. The same process was done with the six
side sections to come up with one side view vector diagram. See figures 5.1. Only those
points with a definite direction were considered in the final analysis. And for those points
with more than one definite direction a resultant direction was assumed. Then a few
vectors were interpolated to complete the picture.
The second step was to bring in the similar views from CFD models from chapter 3
(figures 3.10] and place them side by side with the corresponding vector diagrams in
order to make the comparison process easy. The two summed up vector diagrams are
shown side by side with the CFD vectors diagrams in figures 5.2 and 5.3. And the
interpolated second side view vector diagram is shown in figure 5.4.
127
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
Figure 5.1: Top view and side view showing flow direction at all points before the
final analysis was done.
128
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
129
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
The air flow pattern obtained both by experiment and CFD methods for the other two
geometries i.e., MG-1 and MG-2 are shown in appendix J. The experimental and CFD air
flow patterns look similar in all the three mine geometries. Therefore it can be said with
certainty that CFD can be utilized to solve steady state problems of coalmine scaled-
down model.
Although the results by experiment and CFD methods look similar the air flow at some
points is not exactly the same. This is so because in the experimental method points of
entry for smoke were only at specific locations therefore could not depict air flow
accurately.
As it has already been explained, the objective of this analysis was to investigate the rate
of extraction of initial volume of air out of the test section. Therefore the results for both
130
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
CFD model and experiment were presented in form of time graphs in order to make
comparison process easy. Due to the nature or design of the experiment the results were
presented as a graph of resistance in ohms against time while the numerical results were
presented as a graph of mass fraction of air against time.
To simplify the comparison, the resistance quantities on the y-axis of experimental results
were first converted to mass fraction of air so that both results can be presented on a
single graph. Readings for sensor 2 in table 4.3 is used here to demonstrate the
conversion of experimental readings from ohms to mass fraction of air in the following
steps:
i. The average sensor reading at the scan range where the graph seems to be
flattening was found. This is a point when the smoke is almost completely
extracted out of the test section. For sensor 2 the average reading was found to be
204.488 kohm. This was regarded as the final reading.
ii. The reading at the point when the fan was switched on to begin extraction process
was regarded as the first reading. Extraction began at scan 50 and for sensor 2 the
reading was 180.303 kohm.
iii. The first reading was then subtracted from the final reading. For sensor 2,
iv. Then first reading was subtracted from all the other reading from scan 50 to the
final reading. For sensor 2, scan 50 gives zero and scan 90 is approximately
24.185 kohms.
131
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
v. Finally, solution of step iv was divided by solution of step iii to get experimental
readings in mass fraction of air. For sensor 2, the solution for scan 50 is zero, for
scan 90 it is 1 and for scans in between the solution are fractions.
Each sensor position is compared and analysed separately in order to have a clear picture
of what is happening at specific points. Graphs for sensors 2, 3, 4 and 6 for both CFD
model and experimental model are shown in figures 5.5 to 5.8.
Figure 5.5: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 2.
132
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
Figure 5.6: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 3.
Figure 5.7: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 4.
133
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
Figure 5.8: Comparison of the rate of extraction of initial volume of air by experiment
and CFD methods at location sensor 6.
From the comparison graphs shown in figures 5.4 to 5.7, the rate of air extraction at
heading locations sensor 2,3,4, and 6 can be said to be similar because both plots shows
almost the same pattern and are close to each other. Although the plots at sensor location
6 are not as good as the rest they also look similar. Therefore it can be concluded with
confidence that both CFD and experiment methods produce similar results. In other
words CFD method has been verified to be true therefore further numerical analysis can
be confidently pursued.
As it has already been observed the plots for CFD and experiment results being compared
in the graphs are not exactly the same but are similar. Some of probable causes of
discrepancies between the two results are:
134
CHAPTER 5: COMPARISON: CFD AND EXPERIMENTAL RESULTS
• Friction between air and model wall which exits in the experiment was
disregarded in the CFD analysis.
• In the experiment the process of opening heading inlet and outlet and the
switching-on of the fans did not happen simultaneously but a few seconds after
each other while in CFD everything happens the moment the program starts
running.
• The K-є model was used in numerical analysis. The constants in the model
relations have recommended values for boundary layer calculations. The values
available are, unfortunately, not universal to all fluid problems including
recirculating flows [15]. Therefore, it is possible that the constant used for
modelling was not the right one.
135
CHAPTER 6: FAN POSITON AND OPTIMIZATION
6.1 Introduction
After proving, in chapter 5, that CFD modelling can be utilized to solve ventilation
problems in the heading of scaled down underground coalmine model the next step was
to find the fan positions that result in the shortest time of air extraction for different
heading lengths and widths.
This chapter explains the methodology that was used to find optimum fan positions and
presents a method that can be used to determine the optimum jet fan positions for any
other heading length and width within the concerned heading dimensions.
6.2 Methodology
Five different lengths at mining progression of 7 meters in the actual coalmine heading
were chosen for this exercise. The lengths are 7 m, 14 m, 21 m, 28 m and 35 m and are
equivalent to the following scaled down model: 467 mm, 933 mm, 1400 mm, 1867 mm
and 2333 mm respectively. And two different heading width, a full and half width with
scaled down dimensions 210 mm and 420 mm respectively were also chosen. There were
five different mine geometries for each heading width summing up to ten geometries all
together.
The geometries representing these mine progressions were prepared in CAD as before. In
order to search for the optimum position of the fan for different mine progressions,
several geometries were done for each progression with a jet fan at different positions.
The fan position was only varied along the heading length but its position with respect to
the floor and the side wall remained standard i.e. 0.7m and 1.0 m, respectively.
Therefore the fan position presented in the rest of this dissertation is measured along the
heading length with the entrance of the heading as a datum. See figure 6.1.
136
CHAPTER 6: FAN POSITON AND OPTIMIZATION
After the geometry was prepared it was then exported to Star CCM+ for meshing,
modeling and iteration. To ensure correct results the parameters that were used in the
numerical analysis were exactly the same as the unsteady state analysis in chapter 3. But
unlike in chapter 3 where the presence of smoke was detected by sensors located at a
number of different positions in the heading in this analysis the wall at the blind end of
the heading was used for this purpose. This is because the blind end is the most
problematic area as far as dust and mine gas extraction in the heading is concerned.
Graphs of mass fraction of air against time were plotted for each fan position of a
particular geometry and the fan position that resulted into the shortest extraction time was
considered optimal.
137
CHAPTER 6: FAN POSITON AND OPTIMIZATION
The first geometry to be done was the full heading length, i.e. 2333 mm in the scaled
model, and it is explained in details here to demonstrate how the searching exercise for
optimum fan positions for different heading lengths were accomplished. Initially, the
extraction time for standard position, i.e. 80 mm (1.2 m equivalent in an actual mine)
from heading entrance, for a full width geometry was found. Then the fan was positioned
at 0 mm (right at heading entrance), 500 mm and 1000 mm to get a rough idea of a
probable optimum position of the fan.
6.3 Results
The results for the four position 80 mm, 0 mm, 500 mm and 1000 mm are shown in
figure 6.2 and give the following extraction times in seconds: 165 s, 285 s, 560 s and 765
s, respectively. Please note that the extraction time was read at a mass fraction of air of
0.99. This is the reading were the graph appears to touch gridline 1.0 therefore it is very
easy to see with naked eyes.
138
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Figure 6.2: Graph of mass fraction of air against time showing the time taken to
extracted pollutants at four different fan positions, i.e.; 0 mm, 80 mm, 500
mm and 1000 mm.
It can be observed from these results that optimum position should be at or near 80 mm.
Then two more positions 20 mm behind and beyond 80 mm position, i.e. position 60 mm
and 100 mm, were searched. These gave extraction times of 237 s and 237 s leaving
position 80 mm still optimal. See figures 6.3 and 6.4.
139
CHAPTER 6: FAN POSITON AND OPTIMIZATION
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350
Time {s}
Figure 6.3: Graph of mass fraction of air against time showing that the pollutants were
extracted in 237 seconds at fan position 60 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350
Time {s}
Figure 6.4: Graph of mass fraction of air against time showing that the pollutants were
extracted in 237 seconds at fan position 100 mm.
140
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Then two more positions 10 mm behind and beyond 80 mm position, i.e. position 70 mm
and 90 mm, were tried. These gave extraction times of 192 s and 219 s, again leaving
position 80 mm still optimal. See figures 6.5 and 6.6.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350
Time {s}
Figure 6.5: Graph of mass fraction of air against time showing that the pollutants were
extracted in 192 seconds at fan position 70 mm.
141
CHAPTER 6: FAN POSITON AND OPTIMIZATION
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350
Time {s}
Figure 6.6: Graph of mass fraction of air against time showing that the pollutants were
extracted in 219 seconds at fan position 90 mm.
142
CHAPTER 6: FAN POSITON AND OPTIMIZATION
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300
Time {s}
Figure 6.7: Graph of mass fraction of air against time showing that the pollutants were
extracted in 172 seconds at fan position 75 mm.
Results from the searching exercise for optimum fan position in a full heading length are
summarized in table 6.1 and graph of time against fan position, Lf, figure 6.8. Fan
position 80 mm is an optimum position since it has the shortest extraction time than the
rest.
143
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Table 6.1: A summary of time taken to clear the heading at nine fan positions
800
700
Time to Clear Heading {s}
600
500
400
300
200
100
0 200 400 600 800 1000
Fan Position {mm}
Figure 6.8: A summary of time taken to clear the heading at different fan positions
along the length of the heading.
144
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Looking at graph, figure 6.8, it is observed that when the fan is positioned right at the
entrance (fan position zero) it takes a long time i.e. 285s, to clear the heading. This could
be because of the through-road air flow that is passing almost directly across the fan is
reducing the fan output and consequently, the penetration. See the illustration in figure
6.9a. As the fan is moved away from heading entrance into the heading the effect of this
cross flow is reduced as a result the fan output and penetration increases and time taken
to clear the heading is steadily reduced. This occurs up to fan position of 80mm where
time to clear the heading is 165 seconds. See the illustration in figure 6.9b. From this fan
position onwards time to clear the heading starts to increase. This could be because of the
increased recirculation in the heading. Apart from ventilating the through-road, air
flowing in the through-road also assists to draw out the pollutants out of the heading. As
the fan moves deep into the heading the capacity of through-road air flow to drive the
heading air away is greatly reduced. See the illustration in figure 6.9c.
145
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Figure 6.9: Air patterns at three different fan position to illustrate the difference in
time to completely clear out the test section.
146
CHAPTER 6: FAN POSITON AND OPTIMIZATION
The steps done to find the optimum position of fan in a 2333mm long heading were also
followed to search for optimum position of the fan for the remaining heading lengths.
Since the initial solution gave an indication of the optimum position of the fan even for
the rest of mine geometries therefore the search was narrowed down in the remaining
solutions. The solutions for the searching exercise for heading lengths for two different
widths are given in Appendix L.
The optimum positions of the fan in the heading of scaled down model of the
underground coalmine are summarized in table 6.2.
Table 6.2: Optimum fan position for five different heading lengths and two different
heading widths.
Having found optimum fan position for the five heading lengths at two different widths
the next step was to develop a method that can be used to determine optimum fan
position for any other heading length from 467 to 2333 mm. This method is independent
147
CHAPTER 6: FAN POSITON AND OPTIMIZATION
from CFD. This section presents the method and describes how it was derived and
finally, explains how it can be used.
From the summarized results table a graph of optimum fan position against heading
length was plotted and is shown in figure 6.10. The dimensions or quantities that are in
brackets are for the real mine and are presented for relative reference purposes only.
Figure 6.10: A graph for determining optimum fan position for heading lengths ranging
between 467 mm to 2333 mm. The figure the brackets represent the
equivalent dimensions or measurements in the actual mine.
In the graph, figure 6.10, it is observed that for a short heading length, Lmh, the fan
distance from the entrance, Lf, is small but as Lmh increases Lf also increases taking a
148
CHAPTER 6: FAN POSITON AND OPTIMIZATION
shape similar to j-curve. This occurs between Lmh equal to 467mm and 1400mm. At
1400mm, Lf starts to decrease despite increasing Lmh. At 1867 it starts to increase again
but less rapidly as it did at the beginning.
Demonstration on how the tool is utilized was done using two different examples and
figure 6.9 as follows:
149
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Figure 6.11: A tool for determining optimum fan position with points x and y.
Example 1
Assuming one wants to find out the optimum fan positions at heading length of y for a
half width heading.
A vertical line is drawn from point y1 to intersect half width plot at y2. From y2 a
horizontal line is drawn which intersects the y-axis of the graph at y3.
y3 = Lf = 90mm
But, y1 = Lmh = 1600mm
150
CHAPTER 6: FAN POSITON AND OPTIMIZATION
Therefore, the optimum fan position for heading length of 1300mm is 90mm.
Example 2
Assuming now that the heading width is 75% of full width and it is required to find the
optimum fan position at heading length x1.
Like in the first example, a vertical line is drawn from point x1 to intersect half width. A
midpoint between the intersections of this vertical line with full width and half width
plots is a point where optimum fan position plot for a 75% full heading width is supposed
to pass. This point is marked x2. A horizontal line is then drawn from x2 to intersect y-
axis at x3.
x3 = Lf = 90mm
But, x1 = Lmh = 1300mm
And, 75% of heading width is this case is 315mm
Therefore, the optimum fan position for heading length of 1300mm and width 315mm is
90mm.
151
CHAPTER 7:CONCLUSION AND RECOMMENDATIONS
Briefly, this investigation has highlighted that ARD and methane are main problems
faced in underground coalmines. Further solutions are need and sought that can be
applied to control these concentrations to the allowable levels. Ventilation techniques are
popular methods currently being used for this purpose and was utilized in this study.
Since it is difficult to carry out this kind of investigation in the actual mine comparatively
inexpensive CFD techniques offer an alternative method. Before this technique can be
used it is necessary to verify its application via experiments.
The main objective of this study was to investigate if a CFD modelling technique can
reliably be used to obtain dust and mine gas concentrations in underground coalmine.
Due to practical limitation, CFD modelling for a scaled down underground coalmine
model was verified instead and it is believed that the same idea can be extended to the
actual coalmine.
In the investigation the flow of air was first examined in a steady state condition by CFD
modelling and experimentation on the scaled down mine model. The air flow patterns of
results from the two models were similar. Then the unsteady state condition examination
for the two models was also done to investigate the dilution of mine gas. Like in the
steady state, this also produced similar results as the experimental investigation. Minor
variations did exist as explained in chapter 5. The similarity in both steady and unsteady
state results of the two models means that the CFD modelling technique could be relied
on for further analysis.
After verification of CFD models further numerical analysis was done with an objective
of finding optimum jet fan positions for different heading lengths and widths. The
information found in the analysis was then used to create a graph for determining
optimum fan position for different heading lengths and widths of the scaled down model.
It was observed that for a short heading length, the fan distance from the entrance was
152
CHAPTER 7:CONCLUSION AND RECOMMENDATIONS
small but as heading length was increased fan distance also increased taking a shape
similar to j-curve. This occurred between heading lengths 467mm and 1400mm. At
1400mm, fan distance started to decrease despite increasing heading length. At 1867 it
starts to increase again but less rapidly as it did at the beginning. This could be de to the
effects of through-road air flow on the jet fan and nature of air flow recirculation in the
heading.
During the execution of this investigation, a number of applicable study fields related to
this topic were identified. These are beyond the scope of this investigation. They are as
following:
Although the results of the scaled down model appears convincing the size might
compromise the results of the actual mine. Therefore is necessary to carry out a similar
study on a full scale model rather than just concluding that the model analysis can work
on the actual mine.
In this study the author was not able to explain why the optimum fan positions for
different heading lengths changed in the manner as depicted in the graphs, figure 6.10. A
study can be carried out in order to find out why the graphs behave this way.
This study has proved that CFD modelling can be used to optimize fan position in the
scaled down coalmine model therefore it is possible to extend it to other mine ventilation
and related studies, air conditioning for example.
153
REFERRENCES
REFERRENCES
[1] BELLE B K. 2002. Dust Control for Thick-seam Wall Mines. CSIR –
Miningtek, South Africa.
[8] BOXHO J. 1980. Firedamp Drainage – Handbook for the Coal Mining
Industry in the European Community. Verlag Gluckauf. GMBH. Essen.
Pp. 23 -55.
154
REFERRENCES
[9] FLINT J D. 1990. Mine Gas and Coal Dust Explosions and Methane
Outbursts – Their Causes and Prevention. Dissertation. University of
Witwatersrand, Johannesburg. Pp 3 – 8.
[12] SZLAZAK, N., SZLAZAK, J., TOR, A., OBRACAJ, D., AND
BOROWSKI, M., 2003, Ventilation Systems in Dead-end Headings with
Coal Dust and Methane Hazard. 30th International Conference of Safety in
Mines Research Institutes. South African Institute of Mining and
Metallurgy. Pp 673 – 688.
[15] WHITE Frank M. 2006. Viscous Fluid Flow. Third Edition. Mc Graw Hill
Inc., Singapore. ISBN 007-124493-X. Chapters 1,2 and 3.
155
REFERRENCES
156
REFERRENCES
157
APPENDICES: Appendix A
APPENDICES
Appendix A
The following risk-ranking model (Australian Standards 3931: 1988) as shown in figure
A1 was used for the risk assessment. Table A1 and A2 show the probability and
consequence categories that were used for evaluation purposes.
For the dust explosion (acute) safety hazard, the risk equals the probability of a dust
explosion multiplied by the consequence value (damage and injury costs). For the health
(chronic) hazard, the risk is equal to the time-integrated exposure multiplied by the ‘dose
response’. Exposure equals average concentration multiplied by time exposed (for
example, mg/m³ of respirable dust times years of work). Therefore, reducing the
158
APPENDICES: Appendix A
probability or consequence value can reduce the acute dust explosion hazard risk, and
reducing the concentration or time can reduce the chronic health hazard risk. With the
available data and expert opinion the risk assessment for thick seam wall mining was
derived and is summarized in table A3.
159
APPENDICES: Appendix A
160
APPENDICES: Appendix B
Appendix B
Data presented in appendix B shows explosion fatalities in South Africa coalmines due
methane and combination of methane and coal dust in 100 years (1891 – 1990). The
records show the nature of explosion at a particular colliery, number of deaths involved
and the year in which the accident occurred.
Natal
161
APPENDICES: Appendix B
Ramsay Methane 0
South Africa Methane 2
1904 Ramsay Methane 1
Elandslaagte Methane 0
1905 Crown Methane 0
Glencoe Methane 1
Glencoe Methane 0
Newcastle Stream Methane 0
Dundee Coal Methane 1
South Africa Methane 0
South Africa Methane 1
1906 Elandslaagte Methane 18
Elandslaagte Methane 1
1907 Ramsay Methane 1
1908 Glencoe Methane/Coal Dust 77
Cambrian Methane/Coal Dust 0
Ramsay Methane 1
1909 South Africa Methane 1
1910 Cambrian Methane 0
1911 Utrecht Methane 0
1912 Natal Navigation Methane 0
1914 South Africa Methane 1
1915 Hlobane Methane 1
Hattingspruit Methane 1
1918 Hlobane Methane 0
Northfield Methane 0
1919 Durban Navigation No 2 Methane 1
Northfield Methane 0
1921 Northfield Methane 1
162
APPENDICES: Appendix B
Hlobane Methane 1
Bannockburn Methane 0
Burnside Methane 6
1922 Utrecht Methane 0
Burnside Methane 20
Natal Navigation Methane 0
1923 Hlobane Methane 12
1926 Burnside Methane 1
Durban Navigation No 2 Methane/Coal Dust 125
1928 New Tendege Methane 1
Tshoba Methane 0
1930 Burnside Methane 38
1937 Utrecht Methane 0
1941 Utrecht Methane 15
1943 Northfield Methane/Coal Dust 78
1944 Tshoba Methane 0
Hlobane No 1 Methane 57
1945 Tshoba Methane 0
1948 Cambrian Methane 0
1949 Utrecht Methane 0
1951 Durban Navigation No 2 Methane 9
Northfield Methane 3
1952 Hlobane No 2 Methane/Coal Dust 0
Hlobane No 2 Methane/Coal Dust 1
1953 Northfield Methane 5
1954 Newcatle-Platberg Methane 1
1955 Carnavon Anthracite Methane 1
1956 Cambrian Methane 2
1960 Durban Navigation No 3 Methane 0
163
APPENDICES: Appendix B
164
APPENDICES: Appendix B
165
APPENDICES: Appendix B
166
APPENDICES: Appendix C
Appendix C
167
APPENDICES: Appendix D
Appendix D
168
APPENDICES: Appendix D
169
APPENDICES: Appendix E
Appendix E
Calculations for pressure drop in a pipe with an orifice whose results are indicated in
section 3.2.3.6.
Water flows in a pipe with an orifice as shown in the figure above. From the details given
below find the pressure drop between the inlet and outlet.
Solution
Where subscripts i, o and e stand for inlet, orifice and exit, respectively.
170
APPENDICES: Appendix E
To extract friction factor, f from Moody diagram (appendix D), Reynolds number and
relative roughness values must be found first.
1 × 0.1
Rei = , equation 3.3.
1.007 E − 6
= 1.0E5
f = 0.0158
Therefore, hi = 0.00805 m
And ΔPi = 79 Pa
171
APPENDICES: Appendix F
Appendix F
The readings for the eight sensors at different input voltage values and their best lines of
fit that were used to adjust experimental measured values are as follows:
Sensor 1
172
APPENDICES: Appendix F
Sensor 2
R = 0.9954
100
95
90 Sensor2
85
80
220 240 260 280 300 320
173
APPENDICES: Appendix F
Sensor 3
174
APPENDICES: Appendix F
Sensor 4
175
APPENDICES: Appendix F
Sensor 5
176
APPENDICES: Appendix F
Sensor 6
177
APPENDICES: Appendix F
Sensor 7
178
APPENDICES: Appendix F
Sensor 8
179
APPENDICES: Appendix G
Appendix G
This appendix shows vector diagram developed from video images of MG-1 and MG-2
as seen from the top at levels 1, 2 and 3 and as seen from the side at planes 1 to 6. The
dots are the positions where the smoke enters the test section from the tubes, the arrows
indicate flow direction that were definite and circles around the dots indicate that the flow
direction could not be determined at all and a circle and an arrow indicate that despite the
smoke circulating it was flowing to a particular direction most of the time. Failure to
determine flow direction at particular point was a result of continuous change of smoke
direction during the whole period of observation in other words there was no dominant
direction of flow.
180
APPENDICES: Appendix G
Level 1
Level 2
Level 3
181
APPENDICES: Appendix G
Section 1
Section 2
Section 3
Section 4
Section 5
Section 6
MG2 Level 1
MG2 Level 2
MG2 Level 3
MG-2 AS SEEN FROM THE SIDE
183
APPENDICES: Appendix G
Section 1
Section 2
Section 3
Section 4
Section 5
Section 6
184
APPENDICES: Appendix H
Appendix H
The resistance values measured in unsteady state experiment done on a scaled down model of the underground coalmine. Channels
101 to 108 represent sensors 1 to 8.
185
APPENDICES: Appendix H
16 09/03/2009 15:01:49:431 290389.8 325717.9 361512.1 328326.1 263306 406855.9 264256.1 264844.4
17 09/03/2009 15:01:57:522 290924.7 326184 371425.7 327185.1 264655.7 429065.2 265485.9 265323.9
18 09/03/2009 15:02:05:612 290960.3 325318.3 365472.2 327933.2 265474.8 436719.1 266009.7 265505.9
19 09/03/2009 15:02:13:702 289956.9 325757.8 361538.7 326612.4 264593.5 428272.7 264633.5 265878.8
20 09/03/2009 15:02:21:792 290556.3 326379.4 372236 326319.4 263974.2 432525.9 265874.3 265126.3
21 09/03/2009 15:02:29:882 287666.1 323602.4 350186.7 323948.7 261912 396278.6 263505.8 267086.4
22 09/03/2009 15:02:37:972 288525.1 324199.5 351929.2 324326 262193.9 393665.9 263223.9 261445.9
23 09/03/2009 15:02:46:062 287885.8 323513.6 348839.3 324758.9 261969.7 404327.6 264977.5 260336
24 09/03/2009 15:02:54:153 291743.8 324872.1 361494.4 327107.5 262902 412434.3 265197.3 261714.5
25 09/03/2009 15:03:02:243 288740.5 323760 357212.4 326814.4 261752.2 430199.5 266038.6 261492.5
26 09/03/2009 15:03:10:333 290116.7 325520.3 363587.6 327644.6 256189.3 381909.8 265632.4 261794.4
27 09/03/2009 15:03:18:423 289837 326175.1 347172.2 324938.7 255132.7 372537.8 264928.7 262433.7
28 09/03/2009 15:03:26:513 282001.6 312475 338087.4 313860.4 255534.2 378199.5 262053.6 255442.5
29 09/03/2009 15:03:34:603 281877.8 311235.3 336462.6 313690.6 255689.3 379909.8 262647.4 255744.4
30 09/03/2009 15:03:42:693 281598.1 311190.1 335047.2 313384.7 255132.7 380537.8 261943.7 255383.7
31 09/03/2009 15:03:50:783 276762.4 312210.8 334600.2 312748.4 252622.1 379458.2 261743.9 255286
32 09/03/2009 15:03:58:873 274018.7 306739 334342.4 311116.8 251272.5 379237 262275.6 255664.5
33 09/03/2009 15:04:06:962 272320.6 305359.4 334701.8 311150 249790.7 378376.7 261736.1 254850.1
34 09/03/2009 15:04:15:072 272049.7 304428.2 332914.5 311089.8 244909.5 366457.8 261945.9 255522.4
35 09/03/2009 15:04:23:163 272911 303517.5 331503.3 311121.2 245782.9 365958.4 261889.6 255372.5
36 09/03/2009 15:04:31:254 271201.8 302044.1 328268.4 311112.2 244599.8 356053.6 261820.7 255141.7
37 09/03/2009 15:04:39:365 265929.8 299568.4 324829.9 311451.7 242590.9 355545.3 261461.3 255069.1
38 09/03/2009 15:04:47:475 264258.2 299047 320848.6 310071.5 239787.3 349784.9 261301.8 254787.6
39 09/03/2009 15:04:55:606 263663.4 299817.3 316916.4 309987.3 238832.7 351025.8 261236.6 253957.5
40 09/03/2009 15:05:03:736 263124 296831.6 313564.5 308805.2 236689.5 344366.4 260813.8 253193.8
186
APPENDICES: Appendix H
41 09/03/2009 15:05:11:602 262039.9 296028.1 312669.9 307928.4 235774.9 337691.5 260359.8 253466.9
42 09/03/2009 15:05:19:526 261293.7 293794.9 311886.3 306192.5 235461.8 339238.7 259210.5 252900.8
43 09/03/2009 15:05:27:392 260167.7 292594 308305.8 305062.6 234332.2 337283 256151.7 249899.6
44 09/03/2009 15:05:35:023 258468.2 292394.3 307266.9 305320.1 234290.4 329664.7 255897.6 249564.5
45 09/03/2009 15:05:42:655 258021.1 286323.1 304163.6 303315.7 234169.7 330812.3 253783.3 246452.3
46 09/03/2009 15:05:50:288 257862.8 286596.1 302358.9 303504.3 233079.6 315386.9 251436.8 243297
47 09/03/2009 15:05:57:921 257110 275608.1 288696.1 286762.6 233058.3 310949.5 251152.6 241964.9
48 09/03/2009 15:06:05:554 256755.4 275398.3 287326.5 284600.5 233030.8 311002 251150.4 238952.6
49 09/03/2009 15:06:13:186 256483.3 275492.1 285899.1 284838.1 233673.1 312660.1 250399 239118
50 09/03/2009 15:06:20:818 256347.5 275224.1 288924.7 285654.9 233923.9 318690.1 249987.2 238000.4
51 09/03/2009 15:06:28:450 256461.4 276045.4 287923.6 287224.3 234411.6 320956.5 249754.1 238358.9
52 09/03/2009 15:06:36:082 258289.3 278243 289779.3 287055.6 235009.4 322943.2 250631.8 237514.2
53 09/03/2009 15:06:43:715 260754.4 281341.9 293344.3 288223.2 236132.6 323431.5 250845.7 238655.2
54 09/03/2009 15:06:51:367 260886.4 282434 291874.8 289193.3 236936.2 325997.6 251425.4 238977.1
55 09/03/2009 15:06:58:999 262535.7 282249.8 294072.4 290660.6 240043.9 327857.7 252488.9 243445.5
56 09/03/2009 15:07:06:631 262038.5 283279.8 297994.8 291195.6 240562.6 331751.3 252785.3 244573.2
57 09/03/2009 15:07:14:263 262419.2 285017.9 298070.3 291663.9 241445.3 334399.5 253754.2 246018.3
58 09/03/2009 15:07:21:896 263250.5 285524 301546.5 293588.5 241736.1 337567.1 254375.8 245847.3
59 09/03/2009 15:07:29:840 263499.1 286665 301513.2 292958.1 243498.6 342133.3 255368 247958.4
60 09/03/2009 15:07:37:990 264995.3 289106.7 304731.9 294736.1 242124.7 342637.2 255112.7 247478.9
61 09/03/2009 15:07:46:121 265226.1 289373.1 305448.9 296205.6 243354.5 345482.9 255272.6 248110.5
62 09/03/2009 15:07:54:251 265881 291659.5 306461.1 297162.4 244498.8 347929.2 256533.4 249254.7
63 09/03/2009 15:08:02:382 267849.9 294028 308567.7 298805 245199.1 350195.6 256897.5 249431.9
64 09/03/2009 15:08:10:492 269017.6 295615.2 310716.5 301229 248111.5 353283.3 258489.1 249546.1
65 09/03/2009 15:08:18:603 270686.8 297340 313131.6 302382.3 248325.7 357008.1 259543.5 249365.4
187
APPENDICES: Appendix H
66 09/03/2009 15:08:26:693 270626.9 297788.4 314774.3 302656.4 248567.4 356792.8 259395.8 249833.5
67 09/03/2009 15:08:34:783 270451.5 298108 315833.1 302830.6 248948.6 356767.1 259361.5 250342.4
68 09/03/2009 15:08:42:873 270567.6 298534.2 316943 303072 249414.5 357725.1 259571.2 250470.6
69 09/03/2009 15:08:50:963 271004.5 299397.7 320203.9 303504.3 250038.3 358808.4 259232.7 250801.9
70 09/03/2009 15:08:59:053 271767.9 301320.1 321815.5 305049.3 250888.5 359973.8 259402.5 250656.1
71 09/03/2009 15:09:07:143 273266.2 303084.8 322288.3 305739.7 251518.9 362206.9 259479.4 250278.1
72 09/03/2009 15:09:15:233 273843.4 303133.6 324195.1 306450 252833 364320.2 260187.2 250758.5
73 09/03/2009 15:09:23:323 274651.4 304174.7 325618 307648.7 252915.1 365379 260217.9 250859.3
74 09/03/2009 15:09:31:413 275348.4 305304.6 327642.4 308270.2 253738.7 366961.7 259885.3 250798.6
75 09/03/2009 15:09:39:503 274911.1 305535.5 327333.9 309688.7 253571.1 367012.8 260091.7 250657.6
76 09/03/2009 15:09:47:593 275834.6 305695.3 327928.8 309502.2 254324.7 367119.3 259752.1 250810.2
77 09/03/2009 15:09:55:683 275359.5 306177 327416 309436.7 253809.7 367101.6 260191.2 250910.7
78 09/03/2009 15:10:03:773 275916.7 306523.3 328692.4 309680.9 254088.4 368639.9 260162.8 251146.7
79 09/03/2009 15:10:11:863 275807.9 307049.4 329227.4 309764.9 254150.8 368821.9 260183.2 250959.8
80 09/03/2009 15:10:19:953 275813.4 306191.4 326590.2 309313.5 253962.9 368440.1 260158.7 251078.3
81 09/03/2009 15:10:28:043 275637 306145.9 327926.6 309815.2 253591.1 366875.1 260189.1 250504.5
82 09/03/2009 15:10:36:133 275751.3 306748.6 327438.2 309262.5 254339.1 367445.6 260173.2 250836.1
83 09/03/2009 15:10:44:224 275788.5 307622.1 330761.2 309316.9 254520.1 370258.1 259890.9 251202.3
84 09/03/2009 15:10:52:314 275830.8 307584.3 329953.2 309799.7 254510.7 369563.3 260138.4 251371.3
85 09/03/2009 15:11:00:405 275574.9 308485.6 331769 310495.6 254651.1 370020.7 260221 250799.7
86 09/03/2009 15:11:08:495 275637.1 308525.5 332408.3 310568.9 254671.3 370108.3 260280.9 251274.7
87 09/03/2009 15:11:16:585 275900.4 309158.2 332805.7 310133.8 254703 370060.6 260310.9 251424.5
88 09/03/2009 15:11:24:674 276210.8 308443.4 330983.2 309420.1 254348 370992.9 260405.6 251026.4
89 09/03/2009 15:11:32:764 276829.7 308061.6 331498.2 309555.9 254638.6 369878.5 260481.3 250713.1
90 09/03/2009 15:11:40:855 276294.1 307992.8 331342.8 309356.9 254679.2 370779.8 260436.6 251284.5
188
APPENDICES: Appendix J
Appendix J
A graph of resistance against time plotted from the raw data. In this graph the effects of
switching on and off of heating element can clearly be seen in all the plots. The effect on
plot for sensor 6 has been illustrated; point A being switch on and B switch off.
189
APPENDICES: Appendix K
Appendix K
Steady state experiment and CFD results for mine geometries 1 and 2 are compared in
this appendix as follows:
190
APPENDICES: Appendix K
Air flow pattern appears to be similar in all the three views of MG-1.
191
APPENDICES: Appendix K
192
APPENDICES: Appendix K
Air flow pattern appears to be similar in all the three views of MG-2.
193
APPENDICES: Appendix L
Appendix L
The solutions for the searching exercise for optimum position of jet fan in heading
lengths 467, 933, 1400 and 1867 mm with full heading width are presented in section L.1
of this appendix. And the solutions all the five heading lengths with half heading width
are presented in section L.2. The searching exercise for 2333 mm long heading with full
heading width has already been presented in details in chapter 6. Although more than
three positions were done for each of heading length only the optimum and the nearest
positions behind and beyond the optimum are shown.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1.1: Graph of mass fraction of air against time showing that the pollutants were
extracted in 145 seconds at fan position 10 mm.
194
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.1.2: Graph of mass fraction of air against time showing that the pollutants were
extracted in 108 seconds at fan position 15 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1.3: Graph of mass fraction of air against time showing that the pollutants were
extracted in 148 seconds at fan position 20 mm.
195
APPENDICES: Appendix L
Looking at the results for a 467 mm long heading shown in figures L.1.1 to L.1.3, fan
position 15 mm has the shortest extraction time of 108 seconds as compared to 145 and
148 seconds for fan positions 10 and 20 meters respectively. Therefore fan position 15
mm is optimum for this particular length of heading.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1.4: Graph of mass fraction of air against time showing that the pollutants were
extracted in 141 seconds at fan position 30 mm.
196
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1. 5: Graph of mass fraction of air against time showing that the pollutants were
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1.6: Graph of mass fraction of air against time showing that the pollutants were
extracted in 171 seconds at fan position 40 mm.
197
APPENDICES: Appendix L
For the heading length of 933 mm fan position 35 mm with extraction time of 137
seconds is optimum as compared to positions 30 and 40 mm with extraction time of 141
and 171 seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.1.7: Graph of mass fraction of air against time showing that the pollutants were
198
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200
Time {s}
Figure L.1.8: Graph of mass fraction of air against time showing that the pollutants were
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.1.9: Graph of mass fraction of air against time showing that the pollutants were
For the heading length of 1400 mm fan position 80 mm with extraction time of 174
seconds is optimum as compared to positions 70 and 90 mm with the same extraction
time of 184 seconds.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.1.10: Graph of mass fraction of air against time showing that the pollutants were
200
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.1.11: Graph of mass fraction of air against time showing that the pollutants were
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.1.12: Graph of mass fraction of air against time showing that the pollutants were
201
APPENDICES: Appendix L
For the heading length of 1867 mm fan position 70 mm with extraction time of 184
seconds is optimum as compared to positions 60 and 80 mm with extraction time of 197
and 214 seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.2.1: Graph of mass fraction of air against time showing that the pollutants were
extracted in 113 seconds at fan position 10 mm.
202
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.2.2: Graph of mass fraction of air against time showing that the pollutants were
extracted in 94 seconds at fan position 20 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure J.1.3: Graph of mass fraction of air against time showing that the pollutants were
extracted in 105 seconds at fan position 30 mm.
203
APPENDICES: Appendix L
For the heading length of 467 mm fan position 20 mm with extraction time of 94 seconds
is optimum as compared to positions 10 and 30 mm with extraction time of 113 and 105
seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.2.4: Graph of mass fraction of air against time showing that the pollutants were
extracted in 110 seconds at fan position 40 mm.
204
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.2.5: Graph of mass fraction of air against time showing that the pollutants were
extracted in 91 seconds at fan position 50 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150
Time {s}
Figure L.2.6: Graph of mass fraction of air against time showing that the pollutants were
extracted in 96 seconds at fan position 60 mm.
205
APPENDICES: Appendix L
For the heading length of 933 mm fan position 50 mm with extraction time of 91 seconds
is optimum as compared to positions 40 and 60 mm with extraction time of 110 and 96
seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.2.7: Graph of mass fraction of air against time showing that the pollutants were
extracted in 141 seconds at fan position 100 mm.
206
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250
Time {s}
Figure L.2.8: Graph of mass fraction of air against time showing that the pollutants were
extracted in 136 seconds at fan position 110 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300
Time {s}
Figure L.2.9: Graph of mass fraction of air against time showing that the pollutants were
extracted in 154 seconds at fan position 120 mm.
207
APPENDICES: Appendix L
For the heading length of 1400 mm fan position 110 mm with extraction time of 136
seconds is optimum as compared to positions 100 and 120 mm with extraction time of
141 and 154 seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 100 200 300 400
Time {s}
Figure L.2.10: Graph of mass fraction of air against time showing that the pollutants were
extracted in 282 seconds at fan position 90 mm.
208
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300
Time {s}
Figure L.2.11: Graph of mass fraction of air against time showing that the pollutants were
extracted in 178 seconds at fan position 95 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 100 200 300 400
Time {s}
Figure L.2.12: Graph of mass fraction of air against time showing that the pollutants were
extracted in 216 seconds at fan position 100 mm.
209
APPENDICES: Appendix L
For the heading length of 1867 mm fan position 95 mm with extraction time of 178
seconds is optimum as compared to positions 90 and 100 mm with extraction time of 282
and 216 seconds respectively.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 100 200 300 400
Time {s}
Figure L.2.13: Graph of mass fraction of air against time showing that the pollutants were
extracted in 256 seconds at fan position 90 mm.
210
APPENDICES: Appendix L
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 50 100 150 200 250 300 350
Time {s}
Figure L.2.14: Graph of mass fraction of air against time showing that the pollutants were
extracted in 235 seconds at fan position 100 mm.
1.0
0.8
Mass Fraction of Air
0.6
0.4
0.2
0.0
0 100 200 300 400
Time {s}
Figure L.2.15: Graph of mass fraction of air against time showing that the pollutants were
extracted in 251 seconds at fan position 110 mm.
211
APPENDICES: Appendix L
For the heading length of 2333 mm fan position 100 mm with extraction time of 235
seconds is optimum as compared to positions 90 and 110 mm with extraction time of 256
and 251 seconds respectively.
212